Periodic restart of SolrCloud pods #3087
Labels
compliance
Relating to security compliance or documentation
component/solr-service
Related to Solr-as-a-Service, a brokered Solr offering
component/ssb
Milestone
User Story
In order to implement an alternative to NIST control SI-3, the SSB should do rolling restarts of SolrCloud pods every 15 minutes.
Acceptance Criteria
[ACs should be clearly demoable/verifiable whenever possible. Try specifying them using BDD.]
WHEN I check the ages of all the running SolrCloud pods
THEN I see that no pod is older than 15 minutes
AND I can see that the deployment had rolling restarts about 15 minutes apart.
Background
[Any helpful contextual notes or links to artifacts/evidence, if needed]
See control SI-3 in our SSP.
Security Considerations (required)
This is an alternative to NIST control SI-3. By restarting pods periodically we're ensuring that any compromised pod is only useful for the attacker for a short window of time. An attacker would also have to generate more log traffic to keep compromising pods when they get restarted.
Sketch
[Notes or a checklist reflecting our understanding of the selected approach]
Solr-operator specific solution
More general solution for any EKS workload
kubectl -n {NAMESPACE} rollout restart deploy
.The text was updated successfully, but these errors were encountered: