envio 2.20.1 → 2.21.1

package/src/sources/SourceManager.res

@@ -0,0 +1,467 @@
+open Belt
+
+// Ideally the ChainFetcher name suits this better
+// But currently the ChainFetcher module is immutable
+// and handles both processing and fetching.
+// So this module is to encapsulate the fetching logic only
+// with a mutable state for easier reasoning and testing.
+type t = {
+  sources: Utils.Set.t<Source.t>,
+  maxPartitionConcurrency: int,
+  newBlockFallbackStallTimeout: int,
+  stalledPollingInterval: int,
+  getHeightRetryInterval: (~retry: int) => int,
+  mutable activeSource: Source.t,
+  mutable waitingForNewBlockStateId: option<int>,
+  // Should take into consideration partitions fetching for previous states (before rollback)
+  mutable fetchingPartitionsCount: int,
+}
+
+let getActiveSource = sourceManager => sourceManager.activeSource
+
+let makeGetHeightRetryInterval = (
+  ~initialRetryInterval,
+  ~backoffMultiplicative,
+  ~maxRetryInterval,
+) => {
+  (~retry: int) => {
+    let backoff = if retry === 0 {
+      1
+    } else {
+      retry * backoffMultiplicative
+    }
+    Pervasives.min(initialRetryInterval * backoff, maxRetryInterval)
+  }
+}
+
+let make = (
+  ~sources: array<Source.t>,
+  ~maxPartitionConcurrency,
+  ~newBlockFallbackStallTimeout=20_000,
+  ~stalledPollingInterval=5_000,
+  ~getHeightRetryInterval=makeGetHeightRetryInterval(
+    ~initialRetryInterval=1000,
+    ~backoffMultiplicative=2,
+    ~maxRetryInterval=60_000,
+  ),
+) => {
+  let initialActiveSource = switch sources->Js.Array2.find(source => source.sourceFor === Sync) {
+  | None => Js.Exn.raiseError("Invalid configuration, no data-source for historical sync provided")
+  | Some(source) => source
+  }
+  Prometheus.IndexingMaxConcurrency.set(
+    ~maxConcurrency=maxPartitionConcurrency,
+    ~chainId=initialActiveSource.chain->ChainMap.Chain.toChainId,
+  )
+  Prometheus.IndexingConcurrency.set(
+    ~concurrency=0,
+    ~chainId=initialActiveSource.chain->ChainMap.Chain.toChainId,
+  )
+  {
+    maxPartitionConcurrency,
+    sources: Utils.Set.fromArray(sources),
+    activeSource: initialActiveSource,
+    waitingForNewBlockStateId: None,
+    fetchingPartitionsCount: 0,
+    newBlockFallbackStallTimeout,
+    stalledPollingInterval,
+    getHeightRetryInterval,
+  }
+}
+
+let fetchNext = async (
+  sourceManager: t,
+  ~fetchState: FetchState.t,
+  ~currentBlockHeight,
+  ~executeQuery,
+  ~waitForNewBlock,
+  ~onNewBlock,
+  ~maxPerChainQueueSize,
+  ~stateId,
+) => {
+  let {maxPartitionConcurrency} = sourceManager
+
+  switch fetchState->FetchState.getNextQuery(
+    ~concurrencyLimit={
+      maxPartitionConcurrency - sourceManager.fetchingPartitionsCount
+    },
+    ~maxQueueSize=maxPerChainQueueSize,
+    ~currentBlockHeight,
+    ~stateId,
+  ) {
+  | ReachedMaxConcurrency
+  | NothingToQuery => ()
+  | WaitingForNewBlock =>
+    switch sourceManager.waitingForNewBlockStateId {
+    | Some(waitingStateId) if waitingStateId >= stateId => ()
+    | Some(_) // Case for the prev state before a rollback
+    | None =>
+      sourceManager.waitingForNewBlockStateId = Some(stateId)
+      let currentBlockHeight = await waitForNewBlock(~currentBlockHeight)
+      switch sourceManager.waitingForNewBlockStateId {
+      | Some(waitingStateId) if waitingStateId === stateId => {
+          sourceManager.waitingForNewBlockStateId = None
+          onNewBlock(~currentBlockHeight)
+        }
+      | Some(_) // Don't reset it if we are waiting for another state
+      | None => ()
+      }
+    }
+  | Ready(queries) => {
+      fetchState->FetchState.startFetchingQueries(~queries, ~stateId)
+      sourceManager.fetchingPartitionsCount =
+        sourceManager.fetchingPartitionsCount + queries->Array.length
+      Prometheus.IndexingConcurrency.set(
+        ~concurrency=sourceManager.fetchingPartitionsCount,
+        ~chainId=sourceManager.activeSource.chain->ChainMap.Chain.toChainId,
+      )
+      let _ =
+        await queries
+        ->Array.map(q => {
+          let promise = q->executeQuery
+          let _ = promise->Promise.thenResolve(_ => {
+            sourceManager.fetchingPartitionsCount = sourceManager.fetchingPartitionsCount - 1
+            Prometheus.IndexingConcurrency.set(
+              ~concurrency=sourceManager.fetchingPartitionsCount,
+              ~chainId=sourceManager.activeSource.chain->ChainMap.Chain.toChainId,
+            )
+          })
+          promise
+        })
+        ->Promise.all
+    }
+  }
+}
+
+type status = Active | Stalled | Done
+
+let getSourceNewHeight = async (
+  sourceManager,
+  ~source: Source.t,
+  ~currentBlockHeight,
+  ~status: ref<status>,
+  ~logger,
+) => {
+  let newHeight = ref(0)
+  let retry = ref(0)
+
+  while newHeight.contents <= currentBlockHeight && status.contents !== Done {
+    try {
+      // Use to detect if the source is taking too long to respond
+      let endTimer = Prometheus.SourceGetHeightDuration.startTimer({
+        "source": source.name,
+        "chainId": source.chain->ChainMap.Chain.toChainId,
+      })
+      let height = await source.getHeightOrThrow()
+      endTimer()
+
+      newHeight := height
+      if height <= currentBlockHeight {
+        retry := 0
+        // Slowdown polling when the chain isn't progressing
+        let pollingInterval = if status.contents === Stalled {
+          sourceManager.stalledPollingInterval
+        } else {
+          source.pollingInterval
+        }
+        await Utils.delay(pollingInterval)
+      }
+    } catch {
+    | exn =>
+      let retryInterval = sourceManager.getHeightRetryInterval(~retry=retry.contents)
+      logger->Logging.childTrace({
+        "msg": `Height retrieval from ${source.name} source failed. Retrying in ${retryInterval->Int.toString}ms.`,
+        "source": source.name,
+        "err": exn->Internal.prettifyExn,
+      })
+      retry := retry.contents + 1
+      await Utils.delay(retryInterval)
+    }
+  }
+  Prometheus.SourceHeight.set(
+    ~sourceName=source.name,
+    ~chainId=source.chain->ChainMap.Chain.toChainId,
+    ~blockNumber=newHeight.contents,
+  )
+  newHeight.contents
+}
+
+// Polls for a block height greater than the given block number to ensure a new block is available for indexing.
+let waitForNewBlock = async (sourceManager: t, ~currentBlockHeight) => {
+  let {sources} = sourceManager
+
+  let logger = Logging.createChild(
+    ~params={
+      "chainId": sourceManager.activeSource.chain->ChainMap.Chain.toChainId,
+      "currentBlockHeight": currentBlockHeight,
+    },
+  )
+  logger->Logging.childTrace("Initiating check for new blocks.")
+
+  let syncSources = []
+  let fallbackSources = []
+  sources->Utils.Set.forEach(source => {
+    if (
+      source.sourceFor === Sync ||
+        // Even if the active source is a fallback, still include
+        // it to the list. So we don't wait for a timeout again
+        // if all main sync sources are still not valid
+        source === sourceManager.activeSource
+    ) {
+      syncSources->Array.push(source)
+    } else {
+      fallbackSources->Array.push(source)
+    }
+  })
+
+  let status = ref(Active)
+
+  let (source, newBlockHeight) = await Promise.race(
+    syncSources
+    ->Array.map(async source => {
+      (
+        source,
+        await sourceManager->getSourceNewHeight(~source, ~currentBlockHeight, ~status, ~logger),
+      )
+    })
+    ->Array.concat([
+      Utils.delay(sourceManager.newBlockFallbackStallTimeout)->Promise.then(() => {
+        if status.contents !== Done {
+          status := Stalled
+
+          switch fallbackSources {
+          | [] =>
+            logger->Logging.childWarn(
+              `No new blocks detected within ${(sourceManager.newBlockFallbackStallTimeout / 1000)
+                  ->Int.toString}s. Polling will continue at a reduced rate. For better reliability, refer to our RPC fallback guide: https://docs.envio.dev/docs/HyperIndex/rpc-sync`,
+            )
+          | _ =>
+            logger->Logging.childWarn(
+              `No new blocks detected within ${(sourceManager.newBlockFallbackStallTimeout / 1000)
+                  ->Int.toString}s. Continuing polling with fallback RPC sources from the configuration.`,
+            )
+          }
+        }
+        // Promise.race will be forever pending if fallbackSources is empty
+        // which is good for this use case
+        Promise.race(
+          fallbackSources->Array.map(async source => {
+            (
+              source,
+              await sourceManager->getSourceNewHeight(
+                ~source,
+                ~currentBlockHeight,
+                ~status,
+                ~logger,
+              ),
+            )
+          }),
+        )
+      }),
+    ]),
+  )
+
+  sourceManager.activeSource = source
+
+  // Show a higher level log if we displayed a warning/error after newBlockFallbackStallTimeout
+  let log = status.contents === Stalled ? Logging.childInfo : Logging.childTrace
+  logger->log({
+    "msg": `New blocks successfully found.`,
+    "source": source.name,
+    "newBlockHeight": newBlockHeight,
+  })
+
+  status := Done
+
+  newBlockHeight
+}
+
+let getNextSyncSource = (
+  sourceManager,
+  // This is needed to include the Fallback source to rotation
+  ~initialSource,
+  // After multiple failures start returning fallback sources as well
+  // But don't try it when main sync sources fail because of invalid configuration
+  // note: The logic might be changed in the future
+  ~attemptFallbacks=false,
+) => {
+  let before = []
+  let after = []
+
+  let hasActive = ref(false)
+
+  sourceManager.sources->Utils.Set.forEach(source => {
+    if source === sourceManager.activeSource {
+      hasActive := true
+    } else if (
+      switch source.sourceFor {
+      | Sync => true
+      | Fallback => attemptFallbacks || source === initialSource
+      }
+    ) {
+      (hasActive.contents ? after : before)->Array.push(source)
+    }
+  })
+
+  switch after->Array.get(0) {
+  | Some(s) => s
+  | None =>
+    switch before->Array.get(0) {
+    | Some(s) => s
+    | None => sourceManager.activeSource
+    }
+  }
+}
+
+let executeQuery = async (sourceManager: t, ~query: FetchState.query, ~currentBlockHeight) => {
+  let toBlockRef = ref(
+    switch query.target {
+    | Head => None
+    | EndBlock({toBlock})
+    | Merge({toBlock}) =>
+      Some(toBlock)
+    },
+  )
+  let responseRef = ref(None)
+  let retryRef = ref(0)
+  let initialSource = sourceManager.activeSource
+
+  while responseRef.contents->Option.isNone {
+    let source = sourceManager.activeSource
+    let toBlock = toBlockRef.contents
+    let retry = retryRef.contents
+
+    let logger = Logging.createChild(
+      ~params={
+        "chainId": source.chain->ChainMap.Chain.toChainId,
+        "logType": "Block Range Query",
+        "partitionId": query.partitionId,
+        "source": source.name,
+        "fromBlock": query.fromBlock,
+        "toBlock": toBlock,
+        "addresses": query.addressesByContractName->FetchState.addressesByContractNameCount,
+        "retry": retry,
+      },
+    )
+
+    try {
+      let response = await source.getItemsOrThrow(
+        ~fromBlock=query.fromBlock,
+        ~toBlock,
+        ~addressesByContractName=query.addressesByContractName,
+        ~indexingContracts=query.indexingContracts,
+        ~partitionId=query.partitionId,
+        ~currentBlockHeight,
+        ~selection=query.selection,
+        ~retry,
+        ~logger,
+      )
+      logger->Logging.childTrace({
+        "msg": "Fetched block range from server",
+        "toBlock": response.latestFetchedBlockNumber,
+        "numEvents": response.parsedQueueItems->Array.length,
+        "stats": response.stats,
+      })
+      responseRef := Some(response)
+    } catch {
+    | Source.GetItemsError(error) =>
+      switch error {
+      | UnsupportedSelection(_)
+      | FailedGettingFieldSelection(_)
+      | FailedParsingItems(_) => {
+          let nextSource = sourceManager->getNextSyncSource(~initialSource)
+
+          // These errors are impossible to recover, so we delete the source
+          // from sourceManager so it's not attempted anymore
+          let notAlreadyDeleted = sourceManager.sources->Utils.Set.delete(source)
+
+          // In case there are multiple partitions
+          // failing at the same time. Log only once
+          if notAlreadyDeleted {
+            switch error {
+            | UnsupportedSelection({message}) => logger->Logging.childError(message)
+            | FailedGettingFieldSelection({exn, message, blockNumber, logIndex})
+            | FailedParsingItems({exn, message, blockNumber, logIndex}) =>
+              logger->Logging.childError({
+                "msg": message,
+                "err": exn->Internal.prettifyExn,
+                "blockNumber": blockNumber,
+                "logIndex": logIndex,
+              })
+            | _ => ()
+            }
+          }
+
+          if nextSource === source {
+            %raw(`null`)->ErrorHandling.mkLogAndRaise(
+              ~logger,
+              ~msg="The indexer doesn't have data-sources which can continue fetching. Please, check the error logs or reach out to the Envio team.",
+            )
+          } else {
+            logger->Logging.childInfo({
+              "msg": "Switching to another data-source",
+              "source": nextSource.name,
+            })
+            sourceManager.activeSource = nextSource
+            retryRef := 0
+          }
+        }
+      | FailedGettingItems({attemptedToBlock, retry: WithSuggestedToBlock({toBlock})}) =>
+        logger->Logging.childTrace({
+          "msg": "Failed getting data for the block range. Immediately retrying with the suggested block range from response.",
+          "toBlock": attemptedToBlock,
+          "suggestedToBlock": toBlock,
+        })
+        toBlockRef := Some(toBlock)
+        retryRef := 0
+      | FailedGettingItems({exn, attemptedToBlock, retry: WithBackoff({message, backoffMillis})}) =>
+        // Starting from the 11th failure (retry=10)
+        // include fallback sources for switch
+        // (previously it would consider only sync sources or the initial one)
+        // This is a little bit tricky to find the right number,
+        // because meaning between RPC and HyperSync is different for the error
+        // but since Fallback was initially designed to be used only for height check
+        // just keep the value high
+        let attemptFallbacks = retry >= 10
+
+        let nextSource = switch retry {
+        // Don't attempt a switch on first two failure
+        | 0 | 1 => source
+        | _ =>
+          // Then try to switch every second failure
+          if retry->mod(2) === 0 {
+            sourceManager->getNextSyncSource(~initialSource, ~attemptFallbacks)
+          } else {
+            source
+          }
+        }
+
+        // Start displaying warnings after 4 failures
+        let log = retry >= 4 ? Logging.childWarn : Logging.childTrace
+        logger->log({
+          "msg": message,
+          "toBlock": attemptedToBlock,
+          "backOffMilliseconds": backoffMillis,
+          "retry": retry,
+          "err": exn->Internal.prettifyExn,
+        })
+
+        let shouldSwitch = nextSource !== source
+        if shouldSwitch {
+          logger->Logging.childInfo({
+            "msg": "Switching to another data-source",
+            "source": nextSource.name,
+          })
+          sourceManager.activeSource = nextSource
+        } else {
+          await Utils.delay(Pervasives.min(backoffMillis, 60_000))
+        }
+        retryRef := retryRef.contents + 1
+      }
+    // TODO: Handle more error cases and hang/retry instead of throwing
+    | exn => exn->ErrorHandling.mkLogAndRaise(~logger, ~msg="Failed to fetch block Range")
+    }
+  }
+
+  responseRef.contents->Option.getUnsafe
+}

package/src/sources/SourceManager.resi

@@ -0,0 +1,32 @@
+type t
+
+let make: (
+  ~sources: array<Source.t>,
+  ~maxPartitionConcurrency: int,
+  ~newBlockFallbackStallTimeout: int=?,
+  ~stalledPollingInterval: int=?,
+  ~getHeightRetryInterval: (~retry: int) => int=?,
+) => t
+
+let getActiveSource: t => Source.t
+
+let fetchNext: (
+  t,
+  ~fetchState: FetchState.t,
+  ~currentBlockHeight: int,
+  ~executeQuery: FetchState.query => promise<unit>,
+  ~waitForNewBlock: (~currentBlockHeight: int) => promise<int>,
+  ~onNewBlock: (~currentBlockHeight: int) => unit,
+  ~maxPerChainQueueSize: int,
+  ~stateId: int,
+) => promise<unit>
+
+let waitForNewBlock: (t, ~currentBlockHeight: int) => promise<int>
+
+let executeQuery: (
+  t,
+  ~query: FetchState.query,
+  ~currentBlockHeight: int,
+) => promise<Source.blockRangeFetchResponse>
+
+let makeGetHeightRetryInterval: (~initialRetryInterval: int, ~backoffMultiplicative: int, ~maxRetryInterval: int) => (~retry: int) => int