High Linux load: what evidence should you collect first?
Use load, CPU, memory and I/O together to distinguish compute contention from blocked work.
Start with a specific problem. Find an investigation you can verify.
Use load, CPU, memory and I/O together to distinguish compute contention from blocked work.
Use PostgreSQL logical backups to connect integrity checks, restore drills and RPO / RTO.
Check NS, A/AAAA, CNAME and TTL to distinguish incorrect records, cache differences and connection failures.
Organize investigation around impact, a timeline and tested hypotheses, then make improvements trackable.
Connect termination reasons, previous logs, events and probes before restarting hides the evidence.
Use error logs to distinguish connection and response failures, then check upstream health, connections and timeouts.
Set up OpenVPN on Ubuntu with complete UDP and TCP server and client configurations, covering certificates, forwarding, DNS, full-tunnel egress and certificate revocation.
Calculate P95 correctly for classic histograms and investigate traffic, errors and latency together.
Prepare version history, an observation window and rollback actions for a release you can verify.
Build a password-authenticated remote-access VPN with swanctl, covering certificates, address pools, firewall rules, split routing and DNS, with lessons from maintaining older installations.
Try another keyword or return to all topics.
Searching titles and summaries only