internal/proxy: add initial set of abci metrics (#7115)
This PR adds an initial set of metrics for use ABCI. The initial metrics enable the calculation of timing histograms and call counts for each of the ABCI methods. The metrics are also labeled as either 'sync' or 'async' to determine if the method call was performed using ABCI's `*Async` methods.
An example of these metrics is included here for reference:
```
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0001"} 0
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0004"} 5
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.002"} 12
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.009"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.02"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.1"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.65"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="2"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="6"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="25"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="+Inf"} 13
tendermint_abci_connection_method_timing_sum{chain_id="ci",method="commit",type="sync"} 0.007802058000000001
tendermint_abci_connection_method_timing_count{chain_id="ci",method="commit",type="sync"} 13
```
These metrics can easily be graphed using prometheus's `histogram_quantile(...)` method to pick out a particular quantile to graph or examine. I chose buckets that were somewhat of an estimate of expected range of times for ABCI operations. They start at .0001 seconds and range to 25 seconds. The hope is that this range captures enough possible times to be useful for us and operators.
3 years ago internal/proxy: add initial set of abci metrics (#7115)
This PR adds an initial set of metrics for use ABCI. The initial metrics enable the calculation of timing histograms and call counts for each of the ABCI methods. The metrics are also labeled as either 'sync' or 'async' to determine if the method call was performed using ABCI's `*Async` methods.
An example of these metrics is included here for reference:
```
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0001"} 0
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0004"} 5
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.002"} 12
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.009"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.02"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.1"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.65"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="2"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="6"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="25"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="+Inf"} 13
tendermint_abci_connection_method_timing_sum{chain_id="ci",method="commit",type="sync"} 0.007802058000000001
tendermint_abci_connection_method_timing_count{chain_id="ci",method="commit",type="sync"} 13
```
These metrics can easily be graphed using prometheus's `histogram_quantile(...)` method to pick out a particular quantile to graph or examine. I chose buckets that were somewhat of an estimate of expected range of times for ABCI operations. They start at .0001 seconds and range to 25 seconds. The hope is that this range captures enough possible times to be useful for us and operators.
3 years ago internal/proxy: add initial set of abci metrics (#7115)
This PR adds an initial set of metrics for use ABCI. The initial metrics enable the calculation of timing histograms and call counts for each of the ABCI methods. The metrics are also labeled as either 'sync' or 'async' to determine if the method call was performed using ABCI's `*Async` methods.
An example of these metrics is included here for reference:
```
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0001"} 0
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0004"} 5
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.002"} 12
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.009"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.02"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.1"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.65"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="2"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="6"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="25"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="+Inf"} 13
tendermint_abci_connection_method_timing_sum{chain_id="ci",method="commit",type="sync"} 0.007802058000000001
tendermint_abci_connection_method_timing_count{chain_id="ci",method="commit",type="sync"} 13
```
These metrics can easily be graphed using prometheus's `histogram_quantile(...)` method to pick out a particular quantile to graph or examine. I chose buckets that were somewhat of an estimate of expected range of times for ABCI operations. They start at .0001 seconds and range to 25 seconds. The hope is that this range captures enough possible times to be useful for us and operators.
3 years ago internal/proxy: add initial set of abci metrics (#7115)
This PR adds an initial set of metrics for use ABCI. The initial metrics enable the calculation of timing histograms and call counts for each of the ABCI methods. The metrics are also labeled as either 'sync' or 'async' to determine if the method call was performed using ABCI's `*Async` methods.
An example of these metrics is included here for reference:
```
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0001"} 0
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0004"} 5
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.002"} 12
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.009"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.02"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.1"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.65"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="2"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="6"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="25"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="+Inf"} 13
tendermint_abci_connection_method_timing_sum{chain_id="ci",method="commit",type="sync"} 0.007802058000000001
tendermint_abci_connection_method_timing_count{chain_id="ci",method="commit",type="sync"} 13
```
These metrics can easily be graphed using prometheus's `histogram_quantile(...)` method to pick out a particular quantile to graph or examine. I chose buckets that were somewhat of an estimate of expected range of times for ABCI operations. They start at .0001 seconds and range to 25 seconds. The hope is that this range captures enough possible times to be useful for us and operators.
3 years ago internal/proxy: add initial set of abci metrics (#7115)
This PR adds an initial set of metrics for use ABCI. The initial metrics enable the calculation of timing histograms and call counts for each of the ABCI methods. The metrics are also labeled as either 'sync' or 'async' to determine if the method call was performed using ABCI's `*Async` methods.
An example of these metrics is included here for reference:
```
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0001"} 0
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0004"} 5
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.002"} 12
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.009"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.02"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.1"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.65"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="2"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="6"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="25"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="+Inf"} 13
tendermint_abci_connection_method_timing_sum{chain_id="ci",method="commit",type="sync"} 0.007802058000000001
tendermint_abci_connection_method_timing_count{chain_id="ci",method="commit",type="sync"} 13
```
These metrics can easily be graphed using prometheus's `histogram_quantile(...)` method to pick out a particular quantile to graph or examine. I chose buckets that were somewhat of an estimate of expected range of times for ABCI operations. They start at .0001 seconds and range to 25 seconds. The hope is that this range captures enough possible times to be useful for us and operators.
3 years ago internal/proxy: add initial set of abci metrics (#7115)
This PR adds an initial set of metrics for use ABCI. The initial metrics enable the calculation of timing histograms and call counts for each of the ABCI methods. The metrics are also labeled as either 'sync' or 'async' to determine if the method call was performed using ABCI's `*Async` methods.
An example of these metrics is included here for reference:
```
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0001"} 0
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0004"} 5
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.002"} 12
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.009"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.02"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.1"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.65"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="2"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="6"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="25"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="+Inf"} 13
tendermint_abci_connection_method_timing_sum{chain_id="ci",method="commit",type="sync"} 0.007802058000000001
tendermint_abci_connection_method_timing_count{chain_id="ci",method="commit",type="sync"} 13
```
These metrics can easily be graphed using prometheus's `histogram_quantile(...)` method to pick out a particular quantile to graph or examine. I chose buckets that were somewhat of an estimate of expected range of times for ABCI operations. They start at .0001 seconds and range to 25 seconds. The hope is that this range captures enough possible times to be useful for us and operators.
3 years ago internal/proxy: add initial set of abci metrics (#7115)
This PR adds an initial set of metrics for use ABCI. The initial metrics enable the calculation of timing histograms and call counts for each of the ABCI methods. The metrics are also labeled as either 'sync' or 'async' to determine if the method call was performed using ABCI's `*Async` methods.
An example of these metrics is included here for reference:
```
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0001"} 0
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0004"} 5
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.002"} 12
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.009"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.02"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.1"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.65"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="2"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="6"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="25"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="+Inf"} 13
tendermint_abci_connection_method_timing_sum{chain_id="ci",method="commit",type="sync"} 0.007802058000000001
tendermint_abci_connection_method_timing_count{chain_id="ci",method="commit",type="sync"} 13
```
These metrics can easily be graphed using prometheus's `histogram_quantile(...)` method to pick out a particular quantile to graph or examine. I chose buckets that were somewhat of an estimate of expected range of times for ABCI operations. They start at .0001 seconds and range to 25 seconds. The hope is that this range captures enough possible times to be useful for us and operators.
3 years ago internal/proxy: add initial set of abci metrics (#7115)
This PR adds an initial set of metrics for use ABCI. The initial metrics enable the calculation of timing histograms and call counts for each of the ABCI methods. The metrics are also labeled as either 'sync' or 'async' to determine if the method call was performed using ABCI's `*Async` methods.
An example of these metrics is included here for reference:
```
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0001"} 0
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.0004"} 5
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.002"} 12
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.009"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.02"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.1"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="0.65"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="2"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="6"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="25"} 13
tendermint_abci_connection_method_timing_bucket{chain_id="ci",method="commit",type="sync",le="+Inf"} 13
tendermint_abci_connection_method_timing_sum{chain_id="ci",method="commit",type="sync"} 0.007802058000000001
tendermint_abci_connection_method_timing_count{chain_id="ci",method="commit",type="sync"} 13
```
These metrics can easily be graphed using prometheus's `histogram_quantile(...)` method to pick out a particular quantile to graph or examine. I chose buckets that were somewhat of an estimate of expected range of times for ABCI operations. They start at .0001 seconds and range to 25 seconds. The hope is that this range captures enough possible times to be useful for us and operators.
3 years ago |
|
- package proxy
-
- import (
- "fmt"
- "os"
- "syscall"
-
- abciclient "github.com/tendermint/tendermint/abci/client"
- tmlog "github.com/tendermint/tendermint/libs/log"
- "github.com/tendermint/tendermint/libs/service"
- )
-
- const (
- connConsensus = "consensus"
- connMempool = "mempool"
- connQuery = "query"
- connSnapshot = "snapshot"
- )
-
- // AppConns is the Tendermint's interface to the application that consists of
- // multiple connections.
- type AppConns interface {
- service.Service
-
- // Mempool connection
- Mempool() AppConnMempool
- // Consensus connection
- Consensus() AppConnConsensus
- // Query connection
- Query() AppConnQuery
- // Snapshot connection
- Snapshot() AppConnSnapshot
- }
-
- // NewAppConns calls NewMultiAppConn.
- func NewAppConns(clientCreator abciclient.Creator, metrics *Metrics) AppConns {
- return NewMultiAppConn(clientCreator, metrics)
- }
-
- // multiAppConn implements AppConns.
- //
- // A multiAppConn is made of a few appConns and manages their underlying abci
- // clients.
- // TODO: on app restart, clients must reboot together
- type multiAppConn struct {
- service.BaseService
-
- metrics *Metrics
- consensusConn AppConnConsensus
- mempoolConn AppConnMempool
- queryConn AppConnQuery
- snapshotConn AppConnSnapshot
-
- consensusConnClient abciclient.Client
- mempoolConnClient abciclient.Client
- queryConnClient abciclient.Client
- snapshotConnClient abciclient.Client
-
- clientCreator abciclient.Creator
- }
-
- // NewMultiAppConn makes all necessary abci connections to the application.
- func NewMultiAppConn(clientCreator abciclient.Creator, metrics *Metrics) AppConns {
- multiAppConn := &multiAppConn{
- metrics: metrics,
- clientCreator: clientCreator,
- }
- multiAppConn.BaseService = *service.NewBaseService(nil, "multiAppConn", multiAppConn)
- return multiAppConn
- }
-
- func (app *multiAppConn) Mempool() AppConnMempool {
- return app.mempoolConn
- }
-
- func (app *multiAppConn) Consensus() AppConnConsensus {
- return app.consensusConn
- }
-
- func (app *multiAppConn) Query() AppConnQuery {
- return app.queryConn
- }
-
- func (app *multiAppConn) Snapshot() AppConnSnapshot {
- return app.snapshotConn
- }
-
- func (app *multiAppConn) OnStart() error {
- c, err := app.abciClientFor(connQuery)
- if err != nil {
- return err
- }
- app.queryConnClient = c
- app.queryConn = NewAppConnQuery(c, app.metrics)
-
- c, err = app.abciClientFor(connSnapshot)
- if err != nil {
- app.stopAllClients()
- return err
- }
- app.snapshotConnClient = c
- app.snapshotConn = NewAppConnSnapshot(c, app.metrics)
-
- c, err = app.abciClientFor(connMempool)
- if err != nil {
- app.stopAllClients()
- return err
- }
- app.mempoolConnClient = c
- app.mempoolConn = NewAppConnMempool(c, app.metrics)
-
- c, err = app.abciClientFor(connConsensus)
- if err != nil {
- app.stopAllClients()
- return err
- }
- app.consensusConnClient = c
- app.consensusConn = NewAppConnConsensus(c, app.metrics)
-
- // Kill Tendermint if the ABCI application crashes.
- go app.killTMOnClientError()
-
- return nil
- }
-
- func (app *multiAppConn) OnStop() {
- app.stopAllClients()
- }
-
- func (app *multiAppConn) killTMOnClientError() {
- killFn := func(conn string, err error, logger tmlog.Logger) {
- logger.Error(
- fmt.Sprintf("%s connection terminated. Did the application crash? Please restart tendermint", conn),
- "err", err)
- if killErr := kill(); killErr != nil {
- logger.Error("Failed to kill this process - please do so manually", "err", killErr)
- }
- }
-
- select {
- case <-app.consensusConnClient.Quit():
- if err := app.consensusConnClient.Error(); err != nil {
- killFn(connConsensus, err, app.Logger)
- }
- case <-app.mempoolConnClient.Quit():
- if err := app.mempoolConnClient.Error(); err != nil {
- killFn(connMempool, err, app.Logger)
- }
- case <-app.queryConnClient.Quit():
- if err := app.queryConnClient.Error(); err != nil {
- killFn(connQuery, err, app.Logger)
- }
- case <-app.snapshotConnClient.Quit():
- if err := app.snapshotConnClient.Error(); err != nil {
- killFn(connSnapshot, err, app.Logger)
- }
- }
- }
-
- func (app *multiAppConn) stopAllClients() {
- if app.consensusConnClient != nil {
- if err := app.consensusConnClient.Stop(); err != nil {
- app.Logger.Error("error while stopping consensus client", "error", err)
- }
- }
- if app.mempoolConnClient != nil {
- if err := app.mempoolConnClient.Stop(); err != nil {
- app.Logger.Error("error while stopping mempool client", "error", err)
- }
- }
- if app.queryConnClient != nil {
- if err := app.queryConnClient.Stop(); err != nil {
- app.Logger.Error("error while stopping query client", "error", err)
- }
- }
- if app.snapshotConnClient != nil {
- if err := app.snapshotConnClient.Stop(); err != nil {
- app.Logger.Error("error while stopping snapshot client", "error", err)
- }
- }
- }
-
- func (app *multiAppConn) abciClientFor(conn string) (abciclient.Client, error) {
- c, err := app.clientCreator()
- if err != nil {
- return nil, fmt.Errorf("error creating ABCI client (%s connection): %w", conn, err)
- }
- c.SetLogger(app.Logger.With("module", "abci-client", "connection", conn))
- if err := c.Start(); err != nil {
- return nil, fmt.Errorf("error starting ABCI client (%s connection): %w", conn, err)
- }
- return c, nil
- }
-
- func kill() error {
- p, err := os.FindProcess(os.Getpid())
- if err != nil {
- return err
- }
-
- return p.Signal(syscall.SIGTERM)
- }
|