woodpecker :deciduous_tree: :baby_chick:

Note: This is a weekend experiment project. Use in production, at your own risk.

What?

Woodpecker forks a number of defined workers (2 x num.of cores) + 1 and a master process managing them. When the resident set size increases the defined throttle, master removes the worker from the cluster gracefully, shuts it down, spawn a new fork, adds it back to the cluster.

Why?

Instagram Engineering team wrote a neat article on dismissing garbage colletion in python - https://engineering.instagram.com/dismissing-python-garbage-collection-at-instagram-4dca40b29172 This experiment is to try the same by disabling GC (https://github.com/nodejs/help/issues/462) and relying on the above defined pattern to run the application.

Benchmarks

System Information

• Concurrency: 4
• Number of requests: 100000
• Number of cores: 4 vCPUs per VM
• RAM: 3.6 GB per VM
• Allocating 55150 bytes / 0.4412 MB per request
• RSS throttle size: 80000 bytes / 0.64 per worker

Note: Web server and stress test script were run on different machines (Google Cloud / Virtual)

Default GC config

• Statuses: { '200': 100000 }
• Minimum time to respond: 1ms
• Maximum time to respond: 1942ms
• Average time to respond: 5.5311600000000025ms
• Rate: 612.8089323029973
• Total time: 163183ms

Process watch

Without Mark Sweep

• Statuses: {'200': 100000 }
• Minimum time to respond: 1ms
• Maximum time to respond: 2863ms
• Average time to respond: 5.001519999999979ms
• Rate: 712.6872585771912ms
• Total time: 140314ms

Process Watch