Intermittent long response time because of internal HTTP requests?

I have recently deployed a rust axum web service onto, and while testing (alone, so definitely not heavy usage) I have noticed that the response time becomes incredibly long intermittently (p99 > 15s). For more context on the web service,

  • Memory and CPU utilization looks fine when the response time spiked, so it is probably i/o.
  • The service is hooked to an external postgres database, but I have verified that the database connection/query is not causing the spike.
  • My biggest suspicion is on the internal HTTP requests some of the API calls make to itself (bad design I know, but it is what it is). An API can make multiple internal requests to http://localhost:9000/xxx.
  • This is the configuration I deployed with in fly.toml

internal_port = 9000
force_https = true
auto_stop_machines = true
auto_start_machines = true
min_machines_running = 1
processes = [“app”]
type = “requests”
hard_limit = 1000
soft_limit = 800

interval = “1m0s”
timeout = “5s”
grace_period = “10s”
method = “GET”
path = “/health”

  • I am quite clueless when it comes to networking, could the issue be because it is calling localhost instead of, or because it is making http calls when force_https is set to true?

Thank you for your time :smiley:

1 Like

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.