Sample produced by DevDesk from made-up data. The company, systems and people shown are fictional. Back to DevDesk

Handover pack

3 tickets · 18 systems · 5 notes

Systems catalog

Application / vendor software

Murexprod

What it is / where it runs: murex-prd.corp.local
vendor: Murex
version: 3.1.44
url: https://murex-prd.corp.local
support team: DL-MUREX-SUPPORT
runs on: murex-prd-01..04
install path: /apps/murex
owned by: Trading-Systems
notes: Upstream source of the trade file. Their EOD cutoff is 19:45 - miss it and the position load has nothing to read.
used by: RISK-1423

Autosys job

PRD_RISK_EOD_LOAD_BOXprod

Job name: PRD_RISK_EOD_LOAD_BOX
machine: batch-prd-04.corp.local
job type: BOX
schedule: 20:15 Mon-Fri
downstream: PRD_RISK_EOD_REPORT
owned by: Batch-Ops
notes: Box holds three jobs. If the box goes ON_HOLD nothing alerts - check it first.
used by: RISK-1423

PRD_RISK_EOD_LOAD_STEP1prod

Job name: PRD_RISK_EOD_LOAD_STEP1
machine: batch-prd-04.corp.local
job type: CMD
box job: PRD_RISK_EOD_LOAD_BOX
command: /apps/risk/bin/load_positions.sh $COB
upstream: feed arrival watcher
downstream: PRD_RISK_EOD_LOAD_STEP2
owned by: Batch-Ops
notes: This is the one that fails on month end. Rerun is safe - it truncates staging first.
sendevent -E FORCE_STARTJOB -J PRD_RISK_EOD_LOAD_STEP1
used by: RISK-1423

Person / team to ask

Market Data Support

Name: DL-MDS-SUPPORT
team: Market Data
email: DL-MDS-SUPPORT@corp.local
chat handle: #market-data-support
ask them about: late or missing BOOK_HIER feed
notes: They answer fastest before 16:00. After that raise a ticket in MDS queue.
used by: RISK-1423

Credential location (never the secret)

RISK_APP prod accountprod

Where the credential lives: CyberArk safe RISK-PRD, account RISK_APP
vault: CyberArk
safe: RISK-PRD
account: RISK_APP
request process: raise ACCESS-REQ, needs manager approval, 24h SLA
notes: Never paste the password anywhere - check it out per session.
used by: RISK-1455

Dashboard / monitoring

EOD batch healthprod

URL: https://grafana.corp.local/d/risk-eod/eod-batch
what to watch: load duration panel - normal is 12 min, month end 40 min
alert channel: #risk-batch-alerts
used by: RISK-1423

View / table / procedure

POSITION_STGprod

Object name: RISK_OWNER.POSITION_STG
object type: table
database: RISKPRD
schema: RISK_OWNER
notes: Staging table the loader writes to. Truncated at the start of every run.
used by: RISK-1423

V_TRADE_POSITION_DAILYprod

Object name: RISK_OWNER.V_TRADE_POSITION_DAILY
object type: view
database: RISKPRD
schema: RISK_OWNER
refreshed by: PRD_RISK_EOD_LOAD_BOX
sample query: SELECT * FROM RISK_OWNER.V_TRADE_POSITION_DAILY WHERE COB_DATE = TRUNC(SYSDATE)-1 AND BOOK = :book
notes: Joins POSITION_STG to BOOK_HIER. Month-end rows land late because the hierarchy feed arrives after the position feed.
used by: RISK-1423, RISK-1390

Denodo virtual view

v_position_enrichedprod

Virtual database / view: risk_vdb.v_position_enriched
vdp host: denodo-prd.corp.local
vdp port: 9999
virtual database: risk_vdb
base views: RISKPRD.V_TRADE_POSITION_DAILY, refdata.book_hier
cache mode: full, refreshed 06:00
jdbc url: jdbc:vdb://denodo-prd.corp.local:9999/risk_vdb
owned by: Data-Virtualisation
notes: What the front office actually queries. Cached - so a fix in Oracle is not visible here until the 06:00 refresh.
used by: RISK-1455

Kafka cluster / topic

risk.position.v2prod

Bootstrap servers: kafka-prd-01.corp.local:9092,kafka-prd-02.corp.local:9092,kafka-prd-03.corp.local:9092
topic: risk.position.v2
consumer group: risk-eod-loader
partitions: 12
security protocol: SASL_SSL
schema registry: https://schema-registry-prd.corp.local:8081
retention: 7 days
console command: kafka-console-consumer --bootstrap-server kafka-prd-01:9092 --topic risk.position.v2 --from-beginning --max-messages 5
owned by: Streaming-Platform
notes: Replaces the Tibco feed from Q4. Both run in parallel until then - if numbers disagree, Tibco is still the book of record.
used by: RISK-1423

Log file location

EOD loader logprod

Path: /apps/risk/logs/eod_load_%Y%m%d.log
server: batch-prd-04.corp.local
rotation: kept 30 days, then archived to /nas/risk/logs
grep hint: grep -n 'ORA-\|FATAL' /apps/risk/logs/eod_load_$(date +%Y%m%d).log
access note: read access via group risk-ops-ro
notes: The real error is usually 40 lines above the stack trace.
used by: RISK-1423

Risk API logprod

Path: /apps/risk/api/logs/risk-api.log
server: risk-app-prd-01.corp.local
rotation: hourly
used by: RISK-1455

Oracle DB / connection

RISKPRD - risk warehouseprod

Connection string / TNS alias: jdbc:oracle:thin:@//ora-risk-prd-scan.corp.local:1521/RISKPRD
host: ora-risk-prd-scan.corp.local
port: 1521
service name: RISKPRD
schema: RISK_OWNER
read user: RISK_RO
tns alias: RISKPRD_RO
owned by: DBA-Markets
notes: Read-only user for investigation. Writes go through the batch account only.
DBA raise ticket queue: DBA-MARKETS.
used by: RISK-1423

RISKUAT - risk warehouseuat

Connection string / TNS alias: jdbc:oracle:thin:@//ora-risk-uat.corp.local:1521/RISKUAT
host: ora-risk-uat.corp.local
port: 1521
service name: RISKUAT
schema: RISK_OWNER
owned by: DBA-Markets
notes: Refreshed from prod every Sunday 02:00 - anything you change is gone Monday.
used by: RISK-1390

Code repository / branch

risk/eod-loader

Repo URL: https://git.corp.local/risk/eod-loader
branch: release/2024.09
module: loader-core
build job: jenkins/risk-eod-loader
key files: src/loader/PositionLoader.java, sql/merge_position.sql
used by: RISK-1455

Deployed server / host

batch-prd-04prod

Hostname / IP: batch-prd-04.corp.local
role: Autosys agent + EOD scripts
app path: /apps/risk/bin
access: request via Batch-Ops, read-only by default
owned by: Batch-Ops
used by: RISK-1423

risk-app-prd-01prod

Hostname / IP: risk-app-prd-01.corp.local
role: risk API + scheduler
app path: /apps/risk
deploy command: /apps/risk/bin/deploy.sh <version>
restart command: sudo systemctl restart risk-api
ports: 8443 (api), 9100 (metrics)
access: jump host jump-prd-02, then ssh as svc_risk
owned by: Unix-Midrange
notes: Deploys are blocked 20:00-23:00 while EOD runs.
used by: RISK-1455

Tibco EMS / BW

EMS PRD - risk position busprod

Server URL: tcp://tibco-ems-prd-01.corp.local:7222
host: tibco-ems-prd-01.corp.local
port: 7222
queue: RISK.POSITION.EOD
topic: RISK.POSITION.BROADCAST
jndi url: tibjmsnaming://tibco-ems-prd-01.corp.local:7222
durable name: risk-eod-durable
admin console: https://tibco-ems-prd-01.corp.local:8222
owned by: Middleware
notes: Failover pair with tibco-ems-prd-02 - the client URL must list both or a failover leaves the durable subscriber stuck.
used by: RISK-1423

Access and service requests

Still in flight with other teams. Whoever picks this up needs to keep chasing them - the reference is what the service desk will ask for.

CR-2291approvedoverdue

what: Open 9092 from risk-app-prd-01 to the Kafka brokers
system: risk.position.v2
raised with: Network Services
approver: CAB
team: Network Services
raised: 2026-10-01
needed by: 2026-10-09
chase: Network Services - quiet for 9 days
blocks: RISK-1455
notes: CAB approved. Waiting on the firewall team to schedule.

ACCESS-REQ-88213pending approvaloverdue

what: Read access to RISKPRD for investigation
system: RISKPRD
raised with: IT Service Desk
approver: Line manager - S. Iyer
team: DBA-Markets
raised: 2026-09-29
needed by: 2026-10-06
last chased: 2026-10-04
chase: Line manager - S. Iyer - quiet for 6 days
blocks: RISK-1423
notes: Approval sat in the manager queue while he was on leave. Delegate is P. Rao - chase there next time.

ACCESS-REQ-88540with team

what: Add me to risk-ops-ro on batch-prd-04
system: batch-prd-04
raised with: IT Service Desk
approver: Line manager - S. Iyer
team: Unix-Midrange
raised: 2026-10-05
last chased: 2026-10-09
chase: Unix-Midrange - quiet for 1 days
blocks: RISK-1423
notes: Approved same day. Unix team says next change window is Thursday.

Ask for a masked UAT copy of POSITION_STGto raise

system: RISKUAT
approver: Data Protection Office
team: DBA-Markets
needed by: 2026-10-24
chase: nobody yet - you still have to raise this one
blocks: RISK-1423
notes: Not raised yet - needs a DPO form first, ask Priya which one.

ACCESS-REQ-79115expired

what: Murex read-only login
system: Murex
team: Trading-Systems
raised: 2026-03-24
closed: 2026-09-20
notes: Lapsed after 90 days of no use. Needs raising again if the trade file question comes back.

ACCESS-REQ-81004granted

what: CyberArk access to safe RISK-PRD
system: RISK_APP account
raised with: IT Service Desk
approver: Line manager - S. Iyer
team: IAM
raised: 2026-08-11
closed: 2026-08-23
notes: Took 12 days. Re-request every 6 months - it lapses.

Meetings

2026-10-09 · Incident review - EOD late on the 28thincident

who was there: me, Ravi, Ops manager
chaired by: Batch-Ops
ticket: RISK-1423
discussed: Box was ON_HOLD from the 28th and nobody was paged.
decided: Ops will add an alert for a box left ON_HOLD past 20:30. No code change needed on our side.
actions: Ops: raise the monitoring change.

2026-10-07 · Kafka cutover - parallel run rulesreview

who was there: me, Streaming Platform, Risk BA
chaired by: Streaming Platform
ticket: RISK-1455
discussed: Reviewed the parallel run between Tibco and Kafka.
decided: Tibco stays the book of record until Q4. If the counts disagree, Tibco wins and we raise a defect - do not silently reconcile.
actions: Me: add a count comparison to the loader. Platform: sort the firewall rule (CR-2291).
mine: Add the Tibco vs Kafka count comparison (still open)
follow up on: 2026-10-14

2026-10-01 · Month-end ordering - who fixes the feeddesign

who was there: me, Ravi (Batch-Ops), Market Data Support, S. Iyer
chaired by: S. Iyer
ticket: RISK-1423
discussed: Walked through the month-end failure. Market Data confirmed BOOK_HIER cannot publish before the vendor close file, typically 20:50.
decided: STEP1 will wait on a feed-arrival watcher rather than a fixed time. Market Data will NOT change their publish time. Agreed the report job stays as-is for now.
actions: Ravi: add the watcher job to the box. Market Data: confirm the file pattern. Me: change STEP1 to depend on the watcher.
mine: Change STEP1 to depend on the feed watcher and test in UAT (still open)
follow up on: 2026-10-08

2026-09-19 · Handover of the Denodo layer from the data teamhandover

who was there: me, Data-Virtualisation
chaired by: Data-Virtualisation
discussed: They walked me through the cache refresh and who owns the base views.
decided: We own v_position_enriched from now on; they keep the VDP platform. Cache refresh time is theirs to change, not ours.
actions: Me: document the cache behaviour where the desk can find it.
mine: Document the 06:00 cache refresh behaviour (still open)
follow up on: 2026-10-03

Cheatsheets

IntelliJ IDEA · IDE

Default keymap. macOS column shows it only where it differs.

Also applies to PyCharm, WebStorm, DataGrip and the rest of the JetBrains family - same keymap, different inspections.

Find your way around

Shift Shift: Search everywhere - classes, files, actions, settings, one box
Ctrl+Shift+A / Cmd+Shift+A: Find action - run any menu command by name
Ctrl+N / Cmd+O: Go to class
Ctrl+Shift+N / Cmd+Shift+O: Go to file
Ctrl+Alt+Shift+N / Cmd+Alt+O: Go to symbol
Ctrl+E / Cmd+E: Recent files
Ctrl+Shift+E / Cmd+Shift+E: Recent locations - the one people forget
Ctrl+B / Cmd+B: Go to declaration - Ctrl+Click does the same
Ctrl+Alt+B / Cmd+Alt+B: Go to implementation
Alt+F7 / Alt+F7: Find usages
Ctrl+Alt+Left / Cmd+Alt+Left: Navigate back
Ctrl+Alt+Right / Cmd+Alt+Right: Navigate forward
Ctrl+G / Cmd+L: Go to line
Ctrl+F12 / Cmd+F12: File structure - jump to a method in this file
Alt+1 / Cmd+1: Project tool window
Alt+F12 / Alt+F12: Terminal

Edit

Alt+Enter / Alt+Enter: Show intention actions / quick fix - the single most useful key
Ctrl+Space / Ctrl+Space: Basic code completion
Ctrl+Shift+Enter / Cmd+Shift+Enter: Complete statement - adds the semicolon and braces
Ctrl+P / Cmd+P: Parameter info
Ctrl+Q / F1: Quick documentation
Ctrl+D / Cmd+D: Duplicate line or selection
Ctrl+Y / Cmd+Backspace: Delete line
Ctrl+/ / Cmd+/: Comment line
Ctrl+Shift+/ / Cmd+Alt+/: Block comment
Ctrl+W / Alt+Up: Extend selection
Alt+Shift+Up / Down / Alt+Shift+Up / Down: Move line up / down
Ctrl+Shift+Up / Down / Cmd+Shift+Up / Down: Move statement up / down
Alt+J / Ctrl+G: Add selection for next occurrence - multi-cursor
Ctrl+Alt+Shift+J / Ctrl+Cmd+G: Select all occurrences
Ctrl+Alt+L / Cmd+Alt+L: Reformat code
Ctrl+Alt+O / Ctrl+Alt+O: Optimize imports
Alt+Insert / Cmd+N: Generate - constructor, getters, toString
Ctrl+O / Ctrl+O: Override methods
Ctrl+I / Ctrl+I: Implement methods
Ctrl+Alt+T / Cmd+Alt+T: Surround with - try/catch, if, loop

Refactor

Shift+F6 / Shift+F6: Rename
Ctrl+Alt+Shift+T / Ctrl+T: Refactor this - the whole refactoring menu
Ctrl+Alt+V / Cmd+Alt+V: Extract variable
Ctrl+Alt+M / Cmd+Alt+M: Extract method
Ctrl+Alt+C / Cmd+Alt+C: Extract constant
Ctrl+Alt+F / Cmd+Alt+F: Extract field
F6 / F6: Move

Search and replace

Ctrl+F / Cmd+F: Find in file
Ctrl+R / Cmd+R: Replace in file
Ctrl+Shift+F / Cmd+Shift+F: Find in path
Ctrl+Shift+R / Cmd+Shift+R: Replace in path

Run and debug

Ctrl Ctrl / Ctrl Ctrl: Run anything
Shift+F10 / Ctrl+R: Run
Shift+F9 / Ctrl+D: Debug
Ctrl+F8 / Cmd+F8: Toggle breakpoint
Ctrl+Shift+F8 / Cmd+Shift+F8: View / edit breakpoints - conditional breakpoints live here
F8 / F8: Step over
F7 / F7: Step into
Shift+F8 / Shift+F8: Step out
F9 / Cmd+Alt+R: Resume program
Alt+F8 / Alt+F8: Evaluate expression
Alt+F9 / Alt+F9: Run to cursor

Version control

Ctrl+K / Cmd+K: Commit
Ctrl+Shift+K / Cmd+Shift+K: Push
Ctrl+T / Cmd+T: Update project / pull
Alt+` / Ctrl+V: VCS operations popup
Ctrl+Alt+Z / Cmd+Alt+Z: Rollback / revert
Alt+Shift+C / Alt+Shift+C: Recent changes

Database console (Ultimate / DataGrip)

Ctrl+Enter / Cmd+Enter: Execute statement under the caret
Ctrl+Shift+F10 / Ctrl+Shift+R: Execute the whole script
Ctrl+Alt+Shift+E / Cmd+Alt+Shift+E: Explain plan
Ctrl+F5 / Cmd+F5: Refresh the result set

Visual Studio Code · Editor

Default keybindings. macOS column shows it only where it differs.

Ctrl+K Ctrl+S opens the keyboard shortcut editor - use it when a row here does not match your setup, then fix the row.

Find your way around

Ctrl+Shift+P / Cmd+Shift+P: Command palette - everything is in here
Ctrl+P / Cmd+P: Go to file
Ctrl+Shift+O / Cmd+Shift+O: Go to symbol in file - type : to group by kind
Ctrl+T / Cmd+T: Go to symbol in workspace
Ctrl+G / Ctrl+G: Go to line
F12 / F12: Go to definition
Alt+F12 / Alt+F12: Peek definition - without leaving the file
Shift+F12 / Shift+F12: Find all references
Alt+Left / Ctrl+-: Navigate back
Alt+Right / Ctrl+Shift+-: Navigate forward
Ctrl+Tab / Ctrl+Tab: Switch between open editors

Edit

Ctrl+. / Cmd+.: Quick fix / refactor
F2 / F2: Rename symbol
Ctrl+Space / Ctrl+Space: Trigger suggestion
Ctrl+/ / Cmd+/: Toggle line comment
Shift+Alt+A / Shift+Alt+A: Toggle block comment
Shift+Alt+F / Shift+Alt+F: Format document
Alt+Up / Down / Alt+Up / Down: Move line up / down
Shift+Alt+Up / Down / Shift+Alt+Up / Down: Copy line up / down
Ctrl+Shift+K / Cmd+Shift+K: Delete line
Ctrl+Enter / Cmd+Enter: Insert line below
Ctrl+Shift+Enter / Cmd+Shift+Enter: Insert line above
Ctrl+D / Cmd+D: Add next occurrence to selection - multi-cursor
Ctrl+Shift+L / Cmd+Shift+L: Select all occurrences
Ctrl+Alt+Up / Down / Cmd+Alt+Up / Down: Add cursor above / below
Alt+Click / Alt+Click: Add a cursor wherever you click
Shift+Alt+Right / Ctrl+Shift+Cmd+Right: Expand selection
Ctrl+Shift+\ / Cmd+Shift+\: Jump to matching bracket
Alt+Z / Alt+Z: Toggle word wrap - for the 900-character log line

Search and replace

Ctrl+F / Cmd+F: Find
Ctrl+H / Cmd+Alt+F: Replace
Ctrl+Shift+F / Cmd+Shift+F: Find in files
Ctrl+Shift+H / Cmd+Shift+H: Replace in files
Alt+Enter / Alt+Enter: Select all matches of the current find

Panels and layout

Ctrl+` / Ctrl+`: Toggle terminal
Ctrl+Shift+` / Ctrl+Shift+`: New terminal
Ctrl+B / Cmd+B: Toggle sidebar
Ctrl+Shift+E / Cmd+Shift+E: Explorer
Ctrl+Shift+G / Ctrl+Shift+G: Source control
Ctrl+Shift+D / Cmd+Shift+D: Run and debug
Ctrl+Shift+X / Cmd+Shift+X: Extensions
Ctrl+Shift+M / Cmd+Shift+M: Problems panel
Ctrl+\ / Cmd+\: Split editor
Ctrl+K Z / Cmd+K Z: Zen mode
Ctrl+, / Cmd+,: Settings
Ctrl+K Ctrl+S / Cmd+K Cmd+S: Keyboard shortcuts

Run and debug

F5 / F5: Start debugging
Ctrl+F5 / Ctrl+F5: Run without debugging
F9 / F9: Toggle breakpoint
F10 / F10: Step over
F11 / F11: Step into
Shift+F11 / Shift+F11: Step out
Ctrl+Shift+B / Cmd+Shift+B: Run build task

Visual Studio (C# / VB.NET) · IDE

Windows, default 'General Development Settings' profile.

Careful: this is Microsoft Visual Studio, not VS Code - the keymaps are different and the two get confused constantly. If a shortcut here does nothing, check Tools > Options > Environment > Keyboard: the profile chosen at first launch changes many of these.

Find your way around

Ctrl+T: Go To All - files, types, members, symbols
Ctrl+,: Go To Member / navigate to
F12: Go to definition
Alt+F12: Peek definition
Ctrl+F12: Go to implementation
Shift+F12: Find all references
Ctrl+-: Navigate backward
Ctrl+Shift+-: Navigate forward
Ctrl+Alt+L: Solution Explorer
Ctrl+Alt+O: Output window - where the build actually tells you what broke
Ctrl+\ E: Error List
Ctrl+K Ctrl+K: Toggle bookmark
Ctrl+K Ctrl+N: Next bookmark

Edit

Ctrl+.: Quick actions and refactorings - the equivalent of Alt+Enter
Ctrl+Space: IntelliSense / complete word
Ctrl+Shift+Space: Parameter info
Ctrl+K Ctrl+D: Format document
Ctrl+K Ctrl+F: Format selection
Ctrl+K Ctrl+C: Comment selection
Ctrl+K Ctrl+U: Uncomment selection
Ctrl+K Ctrl+X: Insert snippet
Ctrl+K Ctrl+S: Surround with
Shift+Alt+Up / Down: Add cursor above / below - multi-caret
Ctrl+Alt+Click: Add a cursor
Ctrl+L: Cut line
Alt+Up / Down: Move line up / down
Ctrl+M Ctrl+M: Collapse / expand the current region
Ctrl+M Ctrl+O: Collapse to definitions

Refactor

Ctrl+R Ctrl+R: Rename
Ctrl+R Ctrl+M: Extract method
Ctrl+R Ctrl+E: Encapsulate field
Ctrl+R Ctrl+I: Extract interface
Ctrl+R Ctrl+G: Remove and sort usings

Build, run and debug

Ctrl+Shift+B: Build solution
F5: Start debugging
Ctrl+F5: Start without debugging
Shift+F5: Stop debugging
Ctrl+Shift+F5: Restart debugging
F9: Toggle breakpoint
Ctrl+Shift+F9: Delete all breakpoints
F10: Step over
F11: Step into
Shift+F11: Step out
Ctrl+F10: Run to cursor
Ctrl+Alt+I: Immediate window - evaluate anything mid-break
Ctrl+Alt+C: Call stack
Ctrl+Alt+W 1: Watch window 1
Ctrl+Alt+V L: Locals window
Ctrl+Alt+E: Exception settings - break when a specific exception is thrown

Search and replace

Ctrl+F: Find
Ctrl+H: Replace
Ctrl+Shift+F: Find in files
Ctrl+Shift+H: Replace in files
F3: Find next

C# snippets - type it, then Tab Tab

prop: Auto-implemented property
propfull: Property with a backing field
ctor: Constructor
cw: Console.WriteLine
try: try / catch
tryf: try / finally
foreach: foreach loop
for: for loop
switch: switch statement
#region: Region block
svm: static void Main

VB.NET snippets - type it, then Tab

Sub: Sub procedure
Function: Function
Property: Property
For: For loop
ForEach: For Each loop
Try: Try / Catch block
Select: Select Case block
Class: Class definition

Git · Version control

Same everywhere. Copy the command, fill in your own branch names.

The rule that saves people: anything committed can be recovered with git reflog. Anything never committed cannot. Commit before you experiment.

The daily loop

git status: What is changed, what is staged, which branch you are on
git pull --rebase: Get the latest without a merge bubble
git add -p: Stage hunk by hunk - makes you review your own change first
git commit -m "RISK-1423 fix month-end ordering": Commit with the ticket key in the message
git push: Push the current branch
git push -u origin HEAD: Push a new branch and set its upstream

Branches

git switch -c feature/RISK-1423: Create a branch and switch to it
git switch main: Switch to an existing branch
git branch -vv: Branches with upstream and last commit
git branch -d feature/old: Delete a merged branch
git branch -D feature/old: Force delete an unmerged branch - no warning, be sure
git fetch --prune: Refresh remotes and drop branches deleted on the server
git merge origin/main: Merge the latest main into your branch
git rebase origin/main: Replay your commits on top of main - cleaner history; never on a shared branch

Look at what happened

git log --oneline --graph --decorate -20: Readable recent history
git log -p path/to/file: Every change to one file, with diffs
git log --since=2.weeks --author=you: What have I actually done recently
git diff: Unstaged changes
git diff --staged: Exactly what you are about to commit
git diff main...HEAD: Everything on this branch versus main
git show SHA: One commit in full
git blame -L 40,60 path/to/file: Who wrote these lines, and in which commit
git log -S CONNECTION_STRING: Find the commit that introduced or removed a string - the pickaxe - badly underused

Undo and rescue

git restore path/to/file: Throw away uncommitted changes to a file
git restore --staged path/to/file: Unstage but keep the edit
git commit --amend: Fix the last commit message or add a forgotten file - not after pushing
git reset --soft HEAD~1: Undo the last commit, keep the changes staged
git reset --hard HEAD: Discard everything uncommitted - destructive, no undo
git revert SHA: Undo a commit with a new commit - the safe way on a shared branch
git reflog: Every HEAD you have been at - this is how you recover a lost commit
git checkout SHA -- path/to/file: Restore one file from an old commit

Stash

git stash push -m wip-recon-fix: Park uncommitted work
git stash list: What is parked
git stash pop: Bring the newest one back and drop it
git stash apply stash@{2}: Bring a specific one back and keep it
git stash show -p stash@{0}: See what is inside before applying

Hunting a regression

git bisect start: Begin a binary search for the breaking commit
git bisect bad: Mark the current commit as broken
git bisect good SHA: Mark a known-good commit; git walks you through the rest
git bisect reset: Finish and return to where you were

Setup and config

git config --global user.email you@corp.local: Set your commit identity
git config --list --show-origin: Every setting and which file it came from
git remote -v: Where this repo pushes and pulls from
git clone URL --depth 1: Shallow clone of a very large repo

Unix - getting onto the dev server · Shell

Linux / AIX / Solaris. Run from Git Bash, WSL, PuTTY or a Mac terminal.

Replace the example hosts and paths with your real ones and keep the row. A cheatsheet with your actual hostnames in it is worth ten generic ones.

Get onto the box

ssh user@dev-server.corp.local: Connect
ssh -J user@jump-host user@app-prd-01: Hop through a jump host in one command - -J beats two chained ssh calls
ssh -p 2222 user@host: Connect on a non-default port
ssh-keygen -t ed25519 -C you@corp.local: Create a key pair
ssh-copy-id user@host: Install your public key so it stops asking for a password
cat ~/.ssh/config: Host aliases live here - define a Host block once, then just: ssh dev
exit: Log out - or Ctrl+D

Copy files

scp local.txt user@host:/apps/risk/tmp/: Copy a file TO the server
scp user@host:/apps/risk/logs/eod.log .: Copy a file FROM the server into the current directory
scp -r ./config user@host:/apps/risk/: Copy a whole directory - -r for recursive
scp -P 2222 file user@host:/tmp/: scp on a non-default port - capital P here, lowercase p in ssh
rsync -avz --progress ./build/ user@host:/apps/risk/app/: Copy only what changed, with progress - far better than scp for repeat copies
sftp user@host: Interactive transfer session - then put / get / ls / cd
tar czf logs.tgz /apps/risk/logs/: Bundle a pile of files before copying
tar xzf logs.tgz: Unpack it at the other end

Find your way around

pwd: Where am I
ls -lrt: List by time, newest last - the most-used ls on a log directory
ls -lh: List with human-readable sizes
cd -: Back to the previous directory
find /apps/risk -name *.properties: Find files by name
find /apps/risk -mtime -1 -type f: Files changed in the last day - what did that deploy touch
du -sh *: Size of everything in this directory
df -h: Free disk space - check this first when a batch job fails mysteriously
which java: Which binary is actually being picked up
readlink -f /apps/risk/current: Resolve a symlink to its real path

Logs - the reason you logged in

tail -f /apps/risk/logs/eod.log: Follow a log live - Ctrl+C to stop
tail -200 file.log: Last 200 lines
less file.log: Page through it - slash to search, G for end, g for start, q to quit
grep -n ORA- file.log: Find errors with line numbers
grep -i -C 5 exception file.log: Case-insensitive, with 5 lines of context - context is what you actually need
grep -rn RISKPRD /apps/risk/conf/: Search recursively through config
zgrep ORA- file.log.gz: Grep inside a rotated, gzipped log
wc -l file.log: How many lines
grep -c FATAL file.log: How many matches
tail -f file.log | grep --line-buffered ERROR: Follow, but only show errors - --line-buffered or you see nothing

Processes and ports

ps -ef | grep java: Is it running, and as which user
top: Live CPU and memory - or htop where it exists
kill -15 PID: Ask a process to stop cleanly - always try this before -9
kill -9 PID: Force kill - last resort: no cleanup, no flush, truncated logs
ss -tulpn | grep 8443: What is listening on a port - netstat -tulpn on older boxes
lsof -i :8443: Which process owns a port
lsof -p PID: Which files a process has open - finds the log path you cannot locate
nohup ./run.sh > out.log 2>&1 &: Run something that survives your logout
jobs / fg / bg: Manage background jobs in this shell

Permissions and access

ls -l file: See the permissions
chmod +x script.sh: Make a script executable
chmod 640 file: Owner read-write, group read, others nothing
chown user:group file: Change owner - usually needs sudo
id: Who am I, and which groups am I in - answers most permission-denied questions
sudo -l: What am I allowed to run as root

Staying connected

tmux new -s eod: Start a named session that survives a dropped VPN
tmux attach -t eod: Reattach after you get disconnected
Ctrl+B then D: Detach from tmux and leave it running
screen -S eod / screen -r eod: The older equivalent, on boxes without tmux
watch -n 5 ls -lrt /sftp/inbound: Re-run a command every 5 seconds - for watching a file land

Database from the shell

sqlplus RISK_RO@RISKPRD: Connect to Oracle from the box
sqlplus -S user@db @script.sql: Run a script quietly - never put a password on the command line - it shows in ps
tnsping RISKPRD: Check the TNS alias resolves before blaming the app
echo $ORACLE_HOME ; echo $TNS_ADMIN: Where the client and tnsnames.ora actually are

Tickets

Current sprint: Sprint 42 (2026-10-05 to 2026-10-18)

Sprint 42 - current sprint

RISK-1423 — EOD position load fails on month end

in_progress · high · me · Sprint 42 · autosys,eod,oracle · Jira

What this ticket is about

Month-end runs fail because BOOK_HIER arrives after POSITION_STG is loaded, so the view returns rows with no hierarchy and the report job aborts.
Fix under discussion: make STEP1 wait on the hierarchy feed watcher.

Systems involved

Application / vendor software

Murexprodupstream source

What it is / where it runs: murex-prd.corp.local
vendor: Murex
version: 3.1.44
url: https://murex-prd.corp.local
support team: DL-MUREX-SUPPORT
runs on: murex-prd-01..04
install path: /apps/murex
owned by: Trading-Systems
why here: their 19:45 cutoff drives everything
notes: Upstream source of the trade file. Their EOD cutoff is 19:45 - miss it and the position load has nothing to read.

Autosys job

PRD_RISK_EOD_LOAD_BOXprodparent box

Job name: PRD_RISK_EOD_LOAD_BOX
machine: batch-prd-04.corp.local
job type: BOX
schedule: 20:15 Mon-Fri
downstream: PRD_RISK_EOD_REPORT
owned by: Batch-Ops
why here: check ON_HOLD first
notes: Box holds three jobs. If the box goes ON_HOLD nothing alerts - check it first.

PRD_RISK_EOD_LOAD_STEP1prodthe job that fails

Job name: PRD_RISK_EOD_LOAD_STEP1
machine: batch-prd-04.corp.local
job type: CMD
box job: PRD_RISK_EOD_LOAD_BOX
command: /apps/risk/bin/load_positions.sh $COB
upstream: feed arrival watcher
downstream: PRD_RISK_EOD_LOAD_STEP2
owned by: Batch-Ops
why here: safe to rerun after the feed lands
notes: This is the one that fails on month end. Rerun is safe - it truncates staging first.
sendevent -E FORCE_STARTJOB -J PRD_RISK_EOD_LOAD_STEP1

Person / team to ask

Market Data Supportowns the late feed

Name: DL-MDS-SUPPORT
team: Market Data
email: DL-MDS-SUPPORT@corp.local
chat handle: #market-data-support
ask them about: late or missing BOOK_HIER feed
notes: They answer fastest before 16:00. After that raise a ticket in MDS queue.

Dashboard / monitoring

EOD batch healthprodshows the overrun

URL: https://grafana.corp.local/d/risk-eod/eod-batch
what to watch: load duration panel - normal is 12 min, month end 40 min
alert channel: #risk-batch-alerts

View / table / procedure

POSITION_STGprodwritten by STEP1

Object name: RISK_OWNER.POSITION_STG
object type: table
database: RISKPRD
schema: RISK_OWNER
notes: Staging table the loader writes to. Truncated at the start of every run.

V_TRADE_POSITION_DAILYprodthe view the report reads

Object name: RISK_OWNER.V_TRADE_POSITION_DAILY
object type: view
database: RISKPRD
schema: RISK_OWNER
refreshed by: PRD_RISK_EOD_LOAD_BOX
sample query: SELECT * FROM RISK_OWNER.V_TRADE_POSITION_DAILY WHERE COB_DATE = TRUNC(SYSDATE)-1 AND BOOK = :book
notes: Joins POSITION_STG to BOOK_HIER. Month-end rows land late because the hierarchy feed arrives after the position feed.

Kafka cluster / topic

risk.position.v2prodparallel feed

Bootstrap servers: kafka-prd-01.corp.local:9092,kafka-prd-02.corp.local:9092,kafka-prd-03.corp.local:9092
topic: risk.position.v2
consumer group: risk-eod-loader
partitions: 12
security protocol: SASL_SSL
schema registry: https://schema-registry-prd.corp.local:8081
retention: 7 days
console command: kafka-console-consumer --bootstrap-server kafka-prd-01:9092 --topic risk.position.v2 --from-beginning --max-messages 5
owned by: Streaming-Platform
why here: compare counts against Tibco
notes: Replaces the Tibco feed from Q4. Both run in parallel until then - if numbers disagree, Tibco is still the book of record.

Log file location

EOD loader logprodwhere the error is

Path: /apps/risk/logs/eod_load_%Y%m%d.log
server: batch-prd-04.corp.local
rotation: kept 30 days, then archived to /nas/risk/logs
grep hint: grep -n 'ORA-\|FATAL' /apps/risk/logs/eod_load_$(date +%Y%m%d).log
access note: read access via group risk-ops-ro
notes: The real error is usually 40 lines above the stack trace.

Oracle DB / connection

RISKPRD - risk warehouseprodsource of truth

Connection string / TNS alias: jdbc:oracle:thin:@//ora-risk-prd-scan.corp.local:1521/RISKPRD
host: ora-risk-prd-scan.corp.local
port: 1521
service name: RISKPRD
schema: RISK_OWNER
read user: RISK_RO
tns alias: RISKPRD_RO
owned by: DBA-Markets
why here: query here to prove the rows are missing, not wrong
notes: Read-only user for investigation. Writes go through the batch account only.
DBA raise ticket queue: DBA-MARKETS.

Deployed server / host

batch-prd-04prodwhere the job runs

Hostname / IP: batch-prd-04.corp.local
role: Autosys agent + EOD scripts
app path: /apps/risk/bin
access: request via Batch-Ops, read-only by default
owned by: Batch-Ops

Tibco EMS / BW

EMS PRD - risk position busprodpublishes the positions

Server URL: tcp://tibco-ems-prd-01.corp.local:7222
host: tibco-ems-prd-01.corp.local
port: 7222
queue: RISK.POSITION.EOD
topic: RISK.POSITION.BROADCAST
jndi url: tibjmsnaming://tibco-ems-prd-01.corp.local:7222
durable name: risk-eod-durable
admin console: https://tibco-ems-prd-01.corp.local:8222
owned by: Middleware
why here: durable subscriber stalls after a failover - check before blaming the loader
notes: Failover pair with tibco-ems-prd-02 - the client URL must list both or a failover leaves the durable subscriber stuck.

What happened / what I learned

2026-10-10 15:38 · chat · Ops - Ravi

Box was ON_HOLD from the 28th, nobody was paged. Released it at 21:40.

2026-10-10 15:38 · email · Market Data Support

BOOK_HIER for month end is published after the vendor close file, typically 20:50. It will not move earlier.

2026-10-10 15:38 · self

Reproduced in UAT: run STEP1 before hierarchy load and the view drops ~1,900 rows. Confirms the ordering theory, not a data quality issue.

RISK-1455 — Risk API returns stale positions after restart

open · medium · me · Sprint 42 · api,cache · Jira

What this ticket is about

Cache warms from the previous COB when the API restarts before the EOD load finishes. Needs a readiness check against the load marker table.

Systems involved

Credential location (never the secret)

RISK_APP prod accountprodneeded to restart

Where the credential lives: CyberArk safe RISK-PRD, account RISK_APP
vault: CyberArk
safe: RISK-PRD
account: RISK_APP
request process: raise ACCESS-REQ, needs manager approval, 24h SLA
notes: Never paste the password anywhere - check it out per session.

Denodo virtual view

v_position_enrichedprodwhat the desk actually sees

Virtual database / view: risk_vdb.v_position_enriched
vdp host: denodo-prd.corp.local
vdp port: 9999
virtual database: risk_vdb
base views: RISKPRD.V_TRADE_POSITION_DAILY, refdata.book_hier
cache mode: full, refreshed 06:00
jdbc url: jdbc:vdb://denodo-prd.corp.local:9999/risk_vdb
owned by: Data-Virtualisation
why here: cached until 06:00, so the desk sees stale numbers even after a fix
notes: What the front office actually queries. Cached - so a fix in Oracle is not visible here until the 06:00 refresh.

Log file location

Risk API logprodcache warm messages land here

Path: /apps/risk/api/logs/risk-api.log
server: risk-app-prd-01.corp.local
rotation: hourly

Code repository / branch

risk/eod-loadershares the marker table code

Repo URL: https://git.corp.local/risk/eod-loader
branch: release/2024.09
module: loader-core
build job: jenkins/risk-eod-loader
key files: src/loader/PositionLoader.java, sql/merge_position.sql

Deployed server / host

risk-app-prd-01prodhosts the API

Hostname / IP: risk-app-prd-01.corp.local
role: risk API + scheduler
app path: /apps/risk
deploy command: /apps/risk/bin/deploy.sh <version>
restart command: sudo systemctl restart risk-api
ports: 8443 (api), 9100 (metrics)
access: jump host jump-prd-02, then ssh as svc_risk
owned by: Unix-Midrange
why here: restart window matters
notes: Deploys are blocked 20:00-23:00 while EOD runs.

What happened / what I learned

2026-10-10 15:38 · self

Cache warm happens in @PostConstruct - no check on load state.

Sprint 39

RISK-1390 — Add book hierarchy columns to daily position view

done · low · me · Sprint 39 · oracle

What this ticket is about

Shipped. Kept here because the view definition is the one everyone asks about.

Systems involved

View / table / procedure

V_TRADE_POSITION_DAILYprodchanged here

Object name: RISK_OWNER.V_TRADE_POSITION_DAILY
object type: view
database: RISKPRD
schema: RISK_OWNER
refreshed by: PRD_RISK_EOD_LOAD_BOX
sample query: SELECT * FROM RISK_OWNER.V_TRADE_POSITION_DAILY WHERE COB_DATE = TRUNC(SYSDATE)-1 AND BOOK = :book
notes: Joins POSITION_STG to BOOK_HIER. Month-end rows land late because the hierarchy feed arrives after the position feed.

Oracle DB / connection

RISKUAT - risk warehouseuattested here

Connection string / TNS alias: jdbc:oracle:thin:@//ora-risk-uat.corp.local:1521/RISKUAT
host: ora-risk-uat.corp.local
port: 1521
service name: RISKUAT
schema: RISK_OWNER
owned by: DBA-Markets
notes: Refreshed from prod every Sunday 02:00 - anything you change is gone Monday.

What happened / what I learned

2026-10-10 15:38 · handover

DBA needs 2 days notice for a view change in prod.