Dev.to
7/4/2026

My headless Chrome mirror kept dying on busy mornings. Here is the fix.
Short summary
Headless Chrome startup times exceed 15-second timeout when competing with CPU-intensive content jobs on shared infrastructure. Solution: raise timeout to 45 seconds and add single-shot retry with process cleanup to handle zombie Chrome processes. Acceptable trade-off for hourly background jobs; user-facing flows would require persistent Chrome sockets instead.
- •Cold Chrome launches on busy systems need 45-second deadline, not 15 seconds, to clear realistic worst-case startup windows
- •Add single-retry recovery with process cleanup to handle zombie Chrome processes blocking ports or profile directories
- •Design trade-off: acceptable for background jobs, but user-facing flows should keep Chrome warm with persistent connections
Generated with AI, which can make mistakes.
Is this a good recommendation for you?



