Canister burning through cycles

Developers
45

I apologize for taking so long to follow-up on this thread. We solved the problem a while back and I neglected to close the issue (even after @domwoe was kind enough to remind me). I understand this is not good etiquette on a forum like this.

In short, the cycles leak was due to my failure to initialize the Certified Assets Provenance (CAP) service before deploying our NFT canister. During operation, whenever a token was sold or transferred the CAP service would attempt to report that event using the cronCapEvents method:

  public shared(msg) func cronCapEvents() : async () {
    canistergeekMonitor.collectMetrics();
    logEvent(level_3, "cronCapEvents()");
    var _cont : Bool = true;
    while(_cont){
      var last = List.pop(_capEvents);
      switch(last.0){
        case(?event) {
          _capEvents := last.1;
          try {
            ignore await CapService.insert(event);
          } catch (e) {
            _capEvents := List.push(event, _capEvents);
            logEvent(level_2, "CapService Error : cronCapEvents()");
          };
        };
        case(_) {
          _cont := false;
        };
      };
    };
  };
  public shared(msg) func initCap() : async () {
    canistergeekMonitor.collectMetrics();
    logEvent(level_1, "initCap()");
    if (Option.isNull(capRootBucketId)){
      try {
        capRootBucketId := await CapService.handshake(Principal.toText(Principal.fromActor(this)), 1_000_000_000_000);
      } catch e {};
    };
  };

After executing initCap we no longer experienced any elevated burn rates and have been running smoothly for about a month now. On average we burn ~0.27T cycles/day.

Thank you to everyone who helped with this. It was a great learning experience.

4 Likes