Unexplained cycles burn

Developers
1

Hey all, hoping someone can help me figure out what’s going on here.

I’m losing 2-3 trillion cycles overnight on a backend canister that’s not in production. This has happened a couple times now:

  • Jan 7-8: Balance went from ~5.4T to ~1.8T overnight
  • Jan 9-10: Balance went from ~2.59T to ~365B overnight

That’s roughly 3.6T and 2.2T gone with minimal usage.

What I’ve ruled out:

  • Timer is set to 2hr intervals (not heartbeat)
  • Very light traffic: 169 queries total, 25M instructions
  • No HTTP outcalls
  • Idle burn rate: ~6.87B/day
  • Checked the code for loops/leaks
  • Reserved cycles shows 0

Canister details:

  • ID: f5cpb-qyaaa-aaaah-qdbeq-cai
  • Memory: ~672MB including 1 snapshot
  • Motoko backend

Any pointers? Thanks!

2

It would help if you could share your code, maybe there is a timer triggering one or multiple loops?

3

thanks rem.codes, I fixed it :slight_smile: there was a timer bomb!