Problem
The client's platform was under sustained automated bot load, and had suffered repeated instability over an extended period. Previous remediation attempts had treated each incident as isolated — patching symptoms as they surfaced without ever identifying why the same failure kept returning.
Intervention
A full investigation traced the instability to two compounding causes: resource exhaustion under bot load, and a vulnerable, unoptimised database query in a third-party plugin that automated traffic was repeatedly triggering. The fix addressed both together — WAF rules to filter malicious traffic at the edge, restructured caching to absorb legitimate load, and database indexing to remove the expensive query path entirely.
Result
Since containment, the instability has not recurred. Finding the actual root cause — rather than the next visible symptom — is what turned a recurring incident into a closed one.