Confirm write authority before changing regional traffic
During an outage, reachable does not necessarily mean safe to use. Establish which data path is authoritative before directing customers to it.
Read articleCobnex blogs / Australia
Cobnex blog
AI implementation, software architecture, cloud operations and Australian technology policy.
512 articles
Page 14 of 29
Do not let the next apply undo incident containment
Before resuming deployments, compare emergency cloud changes with the checked-in definition. Preserve the intended protection through a reviewed reconciliation.
Read articleStop exposure before investigating the release in depth
When a candidate is harming users, use the tested control to limit further impact. Preserve enough evidence to understand what already happened.
Read articleA service objective is burning. Find the affected journey
Start with the operations behind the signal, then use infrastructure evidence to locate the cause. The objective tells you about impact, not automatically the failing component.
Read articleA team's cloud cost jumped. Check the basis before the workload
A higher allocated total can come from usage, pricing, currency, credits or changed rules. Separate those causes before asking engineers to remove resources.
Read articleNew connections fail after a rotation
Existing sessions can hide a stale credential until the pool reconnects. Compare target validity, secret version and consumer refresh before rotating again.
Read articleAn invoice export timed out. Establish its outcome first
Before retrying creation, inspect the operation record and the correct Xero organisation. A timeout leaves uncertainty that another request can make worse.
Read articleA CRM value keeps changing back
Repeated reversals usually point to competing writers or a stale full-record update. Trace the field's provenance before making another manual correction.
Read articleThe provider says delivered but the application did nothing
Trace the event from acknowledgement to durable receipt and worker outcome. The missing stage determines whether to replay, repair or reconcile.
Read articleRepeated rate-limit responses need less traffic, not more workers
Identify the shared budget and stop retry amplification before increasing capacity. More concurrency can make a provider throttle harder to recover from.
Read articleStart with one representative failed message
Diagnose a small sample before moving a whole dead-letter queue. Confirm the cause, current business state and effect history before choosing a recovery action.
Read articleA scheduled job ran an hour late
Compare the stored local rule, named zone and resolved instant before changing the server clock. The error may be in interpretation, queueing or display.
Read articleResolve a field conflict without losing the technician's notes
A sync conflict is a disagreement between versions, not a network fault. Preserve the local work, compare the changed fields and record the chosen resolution.
Read articleCorrect the destination without moving the wrong job
Address support needs to identify the role, revision and affected work before applying a correction. A profile edit can otherwise change more than the customer intended.
Read articleInvestigate a one-cent mismatch without editing the balance
Trace the original inputs, rounding policy and provider units before applying a correction. A manual balance change can conceal the defect and break later reversals.
Read articleHelp someone finish a form without making them start again
When a user is blocked, preserve their work and identify the exact interaction failure. A different browser is not a complete diagnosis or a lasting fix.
Read articleFind out whether a missing row is filtered, stale or unavailable
Table support should reconstruct the query and record state before treating a display problem as missing data.
Read articleCheck whether the work exists before asking for another submission
After an expiry complaint, distinguish missing confirmation from missing data. Preserve the draft and verify the original command before retrying.
Read article