Skip to content

agreement: add counters for dropped tasks and slow responses #3861

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
cce merged 1 commit into from
Apr 12, 2022
Merged

agreement: add counters for dropped tasks and slow responses #3861

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions agreement/cryptoVerifier.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,8 +22,12 @@ import (

"github.com/algorand/go-algorand/logging"
"github.com/algorand/go-algorand/protocol"
"github.com/algorand/go-algorand/util/metrics"
)

var voteVerifierOutFullCounter = metrics.MakeCounter(
metrics.MetricName{Name: "algod_agreement_vote_verifier_responses_dropped", Description: "Number of voteVerifier responses dropped due to full channel"})

// TODO put these in config
const (
voteParallelism = 16
Expand Down Expand Up @@ -210,6 +214,7 @@ func (c *poolCryptoVerifier) voteFillWorker(toBundleWait chan<- bundleFuture) {
select {
case c.votes.out <- asyncVerifyVoteResponse{index: votereq.TaskIndex, err: err, cancelled: true}:
default:
voteVerifierOutFullCounter.Inc(nil)
c.log.Infof("poolCryptoVerifier.voteFillWorker unable to write failed enqueue response to output channel")
}
}
Expand Down
9 changes: 9 additions & 0 deletions agreement/pseudonode.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,6 +29,7 @@ import (
"github.com/algorand/go-algorand/logging/logspec"
"github.com/algorand/go-algorand/logging/telemetryspec"
"github.com/algorand/go-algorand/protocol"
"github.com/algorand/go-algorand/util/metrics"
)

// TODO put these in config
Expand All @@ -43,6 +44,9 @@ var errPseudonodeVerifierClosedChannel = errors.New("crypto verifier closed the
var errPseudonodeNoVotes = errors.New("no valid participation keys to generate votes for given round")
var errPseudonodeNoProposals = errors.New("no valid participation keys to generate proposals for given round")

var pseudonodeBacklogFullByType = metrics.NewTagCounter("algod_agreement_pseudonode_tasks_dropped_{TAG}", "Number of pseudonode tasks dropped per type")
var pseudonodeResultTimeoutsByType = metrics.NewTagCounter("algod_agreement_pseudonode_tasks_timeouts_{TAG}", "Number of pseudonode task result timeouts per type")

// A pseudonode creates proposals and votes with a KeyManager which holds participation keys.
//
// It constructs these messages as if they arrived from an external source and were verified.
Expand Down Expand Up @@ -176,6 +180,7 @@ func (n asyncPseudonode) MakeProposals(ctx context.Context, r round, p period) (
return proposalTask.outputChannel(), nil
default:
proposalTask.close()
pseudonodeBacklogFullByType.Add("proposal", 1)
return nil, fmt.Errorf("unable to make proposal for (%d, %d): %w", r, p, errPseudonodeBacklogFull)
}
}
Expand All @@ -193,6 +198,7 @@ func (n asyncPseudonode) MakeVotes(ctx context.Context, r round, p period, s ste
return proposalTask.outputChannel(), nil
default:
proposalTask.close()
pseudonodeBacklogFullByType.Add("vote", 1)
return nil, fmt.Errorf("unable to make vote for (%d, %d, %d): %w", r, p, s, errPseudonodeBacklogFull)
}
}
Expand Down Expand Up @@ -474,6 +480,7 @@ verifiedVotesLoop:
return
case <-outputTimeout:
// we've been waiting for too long for this vote to be written to the output.
pseudonodeResultTimeoutsByType.Add("vote", 1)
t.node.log.Warnf("pseudonode.makeVotes: unable to write vote to output channel for round %d, period %d", t.round, t.period)
outputTimeout = nil
}
Expand Down Expand Up @@ -577,6 +584,7 @@ verifiedVotesLoop:
return
case <-outputTimeout:
// we've been waiting for too long for this vote to be written to the output.
pseudonodeResultTimeoutsByType.Add("pvote", 1)
t.node.log.Warnf("pseudonode.makeProposals: unable to write proposal vote to output channel for round %d, period %d", t.round, t.period)
outputTimeout = nil
}
Expand All @@ -597,6 +605,7 @@ verifiedPayloadsLoop:
return
case <-outputTimeout:
// we've been waiting for too long for this vote to be written to the output.
pseudonodeResultTimeoutsByType.Add("ppayload", 1)
t.node.log.Warnf("pseudonode.makeProposals: unable to write proposal payload to output channel for round %d, period %d", t.round, t.period)
outputTimeout = nil
}
Expand Down