Ideas for
your next move.
Practical thinking for teams building software, securing systems and operating modern infrastructure. Explore our engineering field notes.
550 insights · Page 27 of 46
ALB access logs missing: validate the logging bucket contract
Set up and diagnose Application Load Balancer access logs with the correct S3 Region, supported encryption and delivery policy before relying on request evidence.
Read field note ↗CloudTrail event history is visible but trail delivery is missing
Diagnose AWS CloudTrail trail delivery separately from Event history, checking logging state, S3 permissions and delivery errors without assuming audit retention is healthy.
Read field note ↗Azure Backup alerts: connect protection failures to an actionable notification
Route Azure Backup alerts through reviewed processing rules and action groups, verifying protection state, suppression and response ownership separately.
Read field note ↗Azure Activity Log alerts: monitor the operation and scope you intended
Build Azure Activity Log alerts around the exact operation, status and resource scope, keeping control-plane evidence separate from guest and application health.
Read field note ↗Azure log alerts: align query scope, evaluation windows and alert state
Diagnose Azure log search alerts with the exact query, scope, evaluation window and stateful behavior before changing notification settings or thresholds.
Read field note ↗Azure Monitor Agent: verify DCR association before reinstalling
Diagnose missing Azure VM telemetry through agent health, data collection rule association, source selection and destination evidence rather than repeated reinstalls.
Read field note ↗A failed CloudWatch canary: separate test failure from runner failure
Troubleshoot CloudWatch Synthetics by separating the monitored user journey from canary permissions, VPC connectivity and artifact delivery.
Read field note ↗EventBridge rule DLQs: distinguish target delivery from business completion
Configure and investigate EventBridge rule-target dead-letter delivery, then reconcile business effects before replaying events after a target failure.
Read field note ↗CloudWatch log metric filters: test new events and missing-data semantics
Build CloudWatch log-based metrics with tested event patterns, explicit dimensions and a deliberate no-data policy before attaching production alarms.
Read field note ↗CloudWatch agent metrics: prove memory and disk collection end to end
Set up and diagnose CloudWatch agent memory and disk metrics by checking configuration, identity, endpoint reachability and the exact published metric dimensions.
Read field note ↗A recovery SOP needs business acceptance, not only a started server
Build a disaster-recovery SOP around dependency order, RTO/RPO evidence, controlled cutover and business acceptance rather than treating server startup as recovery.
Read field note ↗Small requests work but transfers stall: investigate path MTU
Diagnose size-dependent network stalls with path-MTU evidence, tunnel overhead and permitted ICMP feedback before changing every interface or masking the symptom.
Read field note ↗