docs: PBTS synchrony issues runbook (#8129)

closes: #7756 

# What does this pull request change?
This pull request adds a new runbook for operators enountering errors related to the new Proposer-Based Timestamps algorithm. The goal of this runbook is to give operators a set of clear steps that they can follow if they are having issues producing blocks because of clock synchronization problems. 
This pull request also renames the `*PrevoteDelay` metrics to drop the term `MessageDelay`. These metrics provide a combined view of `message_delay` + `synchrony` so the name may be confusing.
# Questions to reviewers
* Are there ways to make the set of steps clearer or are there any pieces that seem confusing?
This commit is contained in:
William Banfield
2022-03-17 19:20:46 +00:00
committed by GitHub
parent 02c7199eec
commit 7c91b53999
3 changed files with 227 additions and 11 deletions
+9 -9
View File
@@ -92,12 +92,12 @@ type Metrics struct {
// be above 2/3 of the total voting power of the network defines the endpoint
// the endpoint of the interval. Subtract the proposal timestamp from this endpoint
// to obtain the quorum delay.
QuorumPrevoteMessageDelay metrics.Gauge
QuorumPrevoteDelay metrics.Gauge
// FullPrevoteMessageDelay is the interval in seconds between the proposal
// FullPrevoteDelay is the interval in seconds between the proposal
// timestamp and the timestamp of the latest prevote in a round where 100%
// of the voting power on the network issued prevotes.
FullPrevoteMessageDelay metrics.Gauge
FullPrevoteDelay metrics.Gauge
// ProposalTimestampDifference is the difference between the timestamp in
// the proposal message and the local time of the validator at the time
@@ -257,17 +257,17 @@ func PrometheusMetrics(namespace string, labelsAndValues ...string) *Metrics {
Help: "Time spent per step.",
Buckets: stdprometheus.ExponentialBucketsRange(0.1, 100, 8),
}, append(labels, "step")).With(labelsAndValues...),
QuorumPrevoteMessageDelay: prometheus.NewGaugeFrom(stdprometheus.GaugeOpts{
QuorumPrevoteDelay: prometheus.NewGaugeFrom(stdprometheus.GaugeOpts{
Namespace: namespace,
Subsystem: MetricsSubsystem,
Name: "quorum_prevote_message_delay",
Name: "quorum_prevote_delay",
Help: "Difference in seconds between the proposal timestamp and the timestamp " +
"of the latest prevote that achieved a quorum in the prevote step.",
}, append(labels, "proposer_address")).With(labelsAndValues...),
FullPrevoteMessageDelay: prometheus.NewGaugeFrom(stdprometheus.GaugeOpts{
FullPrevoteDelay: prometheus.NewGaugeFrom(stdprometheus.GaugeOpts{
Namespace: namespace,
Subsystem: MetricsSubsystem,
Name: "full_prevote_message_delay",
Name: "full_prevote_delay",
Help: "Difference in seconds between the proposal timestamp and the timestamp " +
"of the latest prevote that achieved 100% of the voting power in the prevote step.",
}, append(labels, "proposer_address")).With(labelsAndValues...),
@@ -314,8 +314,8 @@ func NopMetrics() *Metrics {
BlockParts: discard.NewCounter(),
BlockGossipReceiveLatency: discard.NewHistogram(),
BlockGossipPartsReceived: discard.NewCounter(),
QuorumPrevoteMessageDelay: discard.NewGauge(),
FullPrevoteMessageDelay: discard.NewGauge(),
QuorumPrevoteDelay: discard.NewGauge(),
FullPrevoteDelay: discard.NewGauge(),
ProposalTimestampDifference: discard.NewHistogram(),
}
}
+2 -2
View File
@@ -2592,12 +2592,12 @@ func (cs *State) calculatePrevoteMessageDelayMetrics() {
_, val := cs.Validators.GetByAddress(v.ValidatorAddress)
votingPowerSeen += val.VotingPower
if votingPowerSeen >= cs.Validators.TotalVotingPower()*2/3+1 {
cs.metrics.QuorumPrevoteMessageDelay.With("proposer_address", cs.Validators.GetProposer().Address.String()).Set(v.Timestamp.Sub(cs.Proposal.Timestamp).Seconds())
cs.metrics.QuorumPrevoteDelay.With("proposer_address", cs.Validators.GetProposer().Address.String()).Set(v.Timestamp.Sub(cs.Proposal.Timestamp).Seconds())
break
}
}
if ps.HasAll() {
cs.metrics.FullPrevoteMessageDelay.With("proposer_address", cs.Validators.GetProposer().Address.String()).Set(pl[len(pl)-1].Timestamp.Sub(cs.Proposal.Timestamp).Seconds())
cs.metrics.FullPrevoteDelay.With("proposer_address", cs.Validators.GetProposer().Address.String()).Set(pl[len(pl)-1].Timestamp.Sub(cs.Proposal.Timestamp).Seconds())
}
}