DIE FRAGE
Why does the same outage keep recurring?
geschlossen · zurück zu Maßnahmen
Ansicht als Gast. Benutzernamen eingeben, um beizutreten und mitzumachen.
Tests that fail intermittently are retried until green
Customers report outages before monitoring does
I mean the pattern over the last two quarters, not a single incident.
Bündel 5diskutieren
The on-call rota has the same two people on it most weeks
Bündel 1in Diskussion (2)
Leadership asks for a root cause within the hour, so the first plausible one wins
I mean the pattern over the last two quarters, not a single incident.
Reliability has no budget line of its own
To be precise: this is about the process, not the person on call that night.
Bündel 4diskutieren
Every team has its own logging format
Bündel 4diskutieren
Success is measured in features shipped per quarter
Dependencies are upgraded only when something breaks
Meaning the repeat outages, not the one-off hardware failure in March.
The service map in the wiki is two years old
Bündel 6diskutieren
Deploys go out on Friday afternoons because that is when the sprint ends
Post-mortem actions are written up and then never scheduled
Bündel 2diskutieren
Rollbacks take longer than the outage they are meant to end
To be precise: this is about the process, not the person on call that night.
Runbooks are out of date within a month of being written
To be precise: this is about the process, not the person on call that night.
Bündel 6in Diskussion (1)
Migrations are run by hand from a laptop
Bündel 3im Fokus ↑ · schließen ✕
Post-mortems assign actions to people who were not in the room
Scope: the services this team runs, not the whole platform.
Feature work always outranks reliability work in planning
Scope: the services this team runs, not the whole platform.
There is no time set aside after an incident, so the write-up is done at midnight
Feature flags are never cleaned up, so nobody knows which paths are live
Meaning the repeat outages, not the one-off hardware failure in March.
Bündel 2diskutieren
Config lives in five places and drifts between them
I mean the pattern over the last two quarters, not a single incident.
Bündel 3diskutieren
Incidents are declared late because declaring one feels like blame
Load tests were last run before the customer base doubled
The same three services cause most of the pages
Bündel 1diskutieren
One engineer knows how the payment path actually works
Scope: the services this team runs, not the whole platform.
Error budgets exist on a slide and nowhere else
The staging environment does not resemble production
The incident channel fills with people asking for status instead of giving it
Nobody owns the shared database schema
To be precise: this is about the process, not the person on call that night.
Retries are unbounded, so a slow dependency becomes a flood
Alerts fire so often that on-call mutes them
I mean the pattern over the last two quarters, not a single incident.
Bündel 5diskutieren
Capacity is added after the outage, never before