You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

202 lines
5.1 KiB

8 years ago
8 years ago
8 years ago
8 years ago
8 years ago
8 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
8 years ago
8 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
8 years ago
8 years ago
8 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
8 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
  1. package proxy
  2. import (
  3. "fmt"
  4. "os"
  5. "syscall"
  6. abciclient "github.com/tendermint/tendermint/abci/client"
  7. tmlog "github.com/tendermint/tendermint/libs/log"
  8. "github.com/tendermint/tendermint/libs/service"
  9. )
  10. const (
  11. connConsensus = "consensus"
  12. connMempool = "mempool"
  13. connQuery = "query"
  14. connSnapshot = "snapshot"
  15. )
  16. // AppConns is the Tendermint's interface to the application that consists of
  17. // multiple connections.
  18. type AppConns interface {
  19. service.Service
  20. // Mempool connection
  21. Mempool() AppConnMempool
  22. // Consensus connection
  23. Consensus() AppConnConsensus
  24. // Query connection
  25. Query() AppConnQuery
  26. // Snapshot connection
  27. Snapshot() AppConnSnapshot
  28. }
  29. // NewAppConns calls NewMultiAppConn.
  30. func NewAppConns(clientCreator abciclient.Creator, metrics *Metrics) AppConns {
  31. return NewMultiAppConn(clientCreator, metrics)
  32. }
  33. // multiAppConn implements AppConns.
  34. //
  35. // A multiAppConn is made of a few appConns and manages their underlying abci
  36. // clients.
  37. // TODO: on app restart, clients must reboot together
  38. type multiAppConn struct {
  39. service.BaseService
  40. metrics *Metrics
  41. consensusConn AppConnConsensus
  42. mempoolConn AppConnMempool
  43. queryConn AppConnQuery
  44. snapshotConn AppConnSnapshot
  45. consensusConnClient abciclient.Client
  46. mempoolConnClient abciclient.Client
  47. queryConnClient abciclient.Client
  48. snapshotConnClient abciclient.Client
  49. clientCreator abciclient.Creator
  50. }
  51. // NewMultiAppConn makes all necessary abci connections to the application.
  52. func NewMultiAppConn(clientCreator abciclient.Creator, metrics *Metrics) AppConns {
  53. multiAppConn := &multiAppConn{
  54. metrics: metrics,
  55. clientCreator: clientCreator,
  56. }
  57. multiAppConn.BaseService = *service.NewBaseService(nil, "multiAppConn", multiAppConn)
  58. return multiAppConn
  59. }
  60. func (app *multiAppConn) Mempool() AppConnMempool {
  61. return app.mempoolConn
  62. }
  63. func (app *multiAppConn) Consensus() AppConnConsensus {
  64. return app.consensusConn
  65. }
  66. func (app *multiAppConn) Query() AppConnQuery {
  67. return app.queryConn
  68. }
  69. func (app *multiAppConn) Snapshot() AppConnSnapshot {
  70. return app.snapshotConn
  71. }
  72. func (app *multiAppConn) OnStart() error {
  73. c, err := app.abciClientFor(connQuery)
  74. if err != nil {
  75. return err
  76. }
  77. app.queryConnClient = c
  78. app.queryConn = NewAppConnQuery(c, app.metrics)
  79. c, err = app.abciClientFor(connSnapshot)
  80. if err != nil {
  81. app.stopAllClients()
  82. return err
  83. }
  84. app.snapshotConnClient = c
  85. app.snapshotConn = NewAppConnSnapshot(c, app.metrics)
  86. c, err = app.abciClientFor(connMempool)
  87. if err != nil {
  88. app.stopAllClients()
  89. return err
  90. }
  91. app.mempoolConnClient = c
  92. app.mempoolConn = NewAppConnMempool(c, app.metrics)
  93. c, err = app.abciClientFor(connConsensus)
  94. if err != nil {
  95. app.stopAllClients()
  96. return err
  97. }
  98. app.consensusConnClient = c
  99. app.consensusConn = NewAppConnConsensus(c, app.metrics)
  100. // Kill Tendermint if the ABCI application crashes.
  101. go app.killTMOnClientError()
  102. return nil
  103. }
  104. func (app *multiAppConn) OnStop() {
  105. app.stopAllClients()
  106. }
  107. func (app *multiAppConn) killTMOnClientError() {
  108. killFn := func(conn string, err error, logger tmlog.Logger) {
  109. logger.Error(
  110. fmt.Sprintf("%s connection terminated. Did the application crash? Please restart tendermint", conn),
  111. "err", err)
  112. if killErr := kill(); killErr != nil {
  113. logger.Error("Failed to kill this process - please do so manually", "err", killErr)
  114. }
  115. }
  116. select {
  117. case <-app.consensusConnClient.Quit():
  118. if err := app.consensusConnClient.Error(); err != nil {
  119. killFn(connConsensus, err, app.Logger)
  120. }
  121. case <-app.mempoolConnClient.Quit():
  122. if err := app.mempoolConnClient.Error(); err != nil {
  123. killFn(connMempool, err, app.Logger)
  124. }
  125. case <-app.queryConnClient.Quit():
  126. if err := app.queryConnClient.Error(); err != nil {
  127. killFn(connQuery, err, app.Logger)
  128. }
  129. case <-app.snapshotConnClient.Quit():
  130. if err := app.snapshotConnClient.Error(); err != nil {
  131. killFn(connSnapshot, err, app.Logger)
  132. }
  133. }
  134. }
  135. func (app *multiAppConn) stopAllClients() {
  136. if app.consensusConnClient != nil {
  137. if err := app.consensusConnClient.Stop(); err != nil {
  138. app.Logger.Error("error while stopping consensus client", "error", err)
  139. }
  140. }
  141. if app.mempoolConnClient != nil {
  142. if err := app.mempoolConnClient.Stop(); err != nil {
  143. app.Logger.Error("error while stopping mempool client", "error", err)
  144. }
  145. }
  146. if app.queryConnClient != nil {
  147. if err := app.queryConnClient.Stop(); err != nil {
  148. app.Logger.Error("error while stopping query client", "error", err)
  149. }
  150. }
  151. if app.snapshotConnClient != nil {
  152. if err := app.snapshotConnClient.Stop(); err != nil {
  153. app.Logger.Error("error while stopping snapshot client", "error", err)
  154. }
  155. }
  156. }
  157. func (app *multiAppConn) abciClientFor(conn string) (abciclient.Client, error) {
  158. c, err := app.clientCreator()
  159. if err != nil {
  160. return nil, fmt.Errorf("error creating ABCI client (%s connection): %w", conn, err)
  161. }
  162. c.SetLogger(app.Logger.With("module", "abci-client", "connection", conn))
  163. if err := c.Start(); err != nil {
  164. return nil, fmt.Errorf("error starting ABCI client (%s connection): %w", conn, err)
  165. }
  166. return c, nil
  167. }
  168. func kill() error {
  169. p, err := os.FindProcess(os.Getpid())
  170. if err != nil {
  171. return err
  172. }
  173. return p.Signal(syscall.SIGTERM)
  174. }