Incidents/2023-05-23 wdqs CODFW 5xx errors

document status: draft

Summary

Incident metadata (see Incident Scorecard)

Incident ID	2023-05-23 wdqs CODFW 5xx errors	Start	2023-05-23 10:48:00
Task	T337327	End	2023-05-23 19:24:00
People paged		Responder count	5
Coordinators	Brian King	Affected metrics/SLOs	WDQS SLO
Impact	A subset of users accessing the WDQS service from our CODFW datacenter received errors and timeouts.

An expensive query sent over and over again by an external user(s) caused errors and timeouts for users accessing the WDQS service from our CODFW datacenter. A requestctl rule was put in place to mitigate the issue.

Timeline

Link to a specific offset in SAL using the SAL tool at https://sal.toolforge.org/ (example)

All times in UTC

• 1047 first Icinga alerts fire. Example verbiage: - PyBal backends health check on lvs2010 is CRITICAL: PYBAL CRITICAL - CRITICAL - wdqs-heavy-queries_8888: Servers wdqs2009.codfw.wmnet are marked down but pooled
• 1536 Rolling restart of the WDQS service in CODFW temporarily stabilizes the service
• 1630 Time spent in old garbage collection (metric that closely correlates with the outage) starts to rise again (see graph).
• 1821 Requestctl rule to mitigate abuse deployed, followed by a rolling restart of the WDQS service in CODFW. Service starts to stabilize.
• 1924 Old GC is back down to the levels of the non-affected datacenter (eqiad). Incident closed.

Detection

Write how the issue was first detected. Was automated monitoring first to detect it? Or a human reporting an error

Detected by monitoring

Copy the relevant alerts that fired in this section.

See above for example.

Did the appropriate alert(s) fire?

Yes

Was the alert volume manageable?

Yes

Did they point to the problem with as much accuracy as possible?

No. We saw increased 5xx errors and lots of time spent in old garbage collection, but it's still difficult to troubleshoot this type of abuse.

Conclusions

What went well?

• Alerts fired promptly
• Other SREs provided lots of help.

What went poorly?

• Finding the right query/abusive IP etc for WDQS takes a long time, and solutions are often incomplete due to the fragile nature of the service.

Where did we get lucky?

• SMEs for the service were at work when the alerts fired

OPTIONAL: (Use bullet points) for example: user's error report was exceptionally detailed, incident occurred when the most people were online to assist, etc

Links to relevant documentation

• …

Add links to information that someone responding to this alert should have (runbook, plus supporting docs). If that documentation does not exist, add an action item to create it.

Actionables

• …

Create a list of action items that will help prevent this from happening again as much as possible. Link to or create a Phabricator task for every step.

Add the #Sustainability (Incident Followup) and the #SRE-OnFire Phabricator tag to these tasks.

Scorecard

Incident Engagement ScoreCard

	Question	Answer (yes/no)	Notes
People	Were the people responding to this incident sufficiently different than the previous five incidents?
	Were the people who responded prepared enough to respond effectively
	Were fewer than five people paged?
	Were pages routed to the correct sub-team(s)?
	Were pages routed to online (business hours) engineers?  Answer “no” if engineers were paged after business hours.
Process	Was the "Incident status" section atop the Google Doc kept up-to-date during the incident?
	Was a public wikimediastatus.net entry created?
	Is there a phabricator task for the incident?
	Are the documented action items assigned?
	Is this incident sufficiently different from earlier incidents so as not to be a repeat occurrence?
Tooling	To the best of your knowledge was the open task queue free of any tasks that would have prevented this incident? Answer “no” if there are open tasks that would prevent this incident or make mitigation easier if implemented.
	Were the people responding able to communicate effectively during the incident with the existing tooling?
	Did existing monitoring notify the initial responders?
	Were the engineering tools that were to be used during the incident, available and in service?
	Were the steps taken to mitigate guided by an existing runbook?
Total score (count of all “yes” answers above)