Job Types

This guide outlines different job types.

Solidity cron jobs

Executes a job on a schedule. Does not rely on any kind of external trigger.

Spec format

type            = "cron"
schemaVersion   = 1
evmChainID      = 1
schedule        = "CRON_TZ=UTC * */20 * * * *"
externalJobID       = "0EEC7E1D-D0D2-476C-A1A8-72DFB6633F01"
observationSource   = """
    fetch    [type="http" method=GET url="https://chain.link/ETH-USD"]
    parse    [type="jsonparse" path="data,price"]
    multiply [type="multiply" times=100]

    fetch -> parse -> multiply
"""

Unique fields

schedule: the frequency with which the job is to be run. There are two ways to specify this:
- Traditional UNIX cron format, but with 6 fields, not 5. The extra field allows for "seconds" granularity. Note: you must specify the CRON_TZ=... parameter if you use this format.
- @ shorthand, e.g. @every 1h. This shorthand does not take account of the node's timezone, rather, it simply begins counting down the moment that the job is added to the node (or the node is rebooted). As such, no CRON_TZ parameter is needed.

For all supported schedules, please refer to the cron library documentation.

Job type specific pipeline variables

$(jobSpec.databaseID): the ID of the job spec in the local database. You shouldn't need this in 99% of cases.
$(jobSpec.externalJobID): the globally-unique job ID for this job. Used to coordinate between node operators in certain cases.
$(jobSpec.name): the local name of the job.
$(jobRun.meta): a map of metadata that can be sent to a bridge, etc.

Executes a job upon receipt of an explicit request made by a user. The request is detected via a log emitted by an Oracle or Operator contract. This is similar to the legacy ethlog/runlog style of jobs.

Spec format

type                = "directrequest"
schemaVersion       = 1
evmChainID          = 1
name                = "example eth request event spec"
contractAddress     = "0x613a38AC1659769640aaE063C651F48E0250454C"

# Optional fields:
# requesters        = [
#   "0xAaAA1F8ee20f5565510b84f9353F1E333e753B7a",
#   "0xBbBb70f0E81c6F3430dfDc9fa02fB22bDD818c4E"
# ]
# minContractPaymentLinkJuels = "100000000000000"
# externalJobID = "0EEC7E1D-D0D2-476C-A1A8-72DFB6633F02"
# minIncomingConfirmations = 10

observationSource   = """
    ds          [type="http" method=GET url="http://example.com"]
    ds_parse    [type="jsonparse" path="USD"]
    ds_multiply [type="multiply" times=100]

    ds -> ds_parse -> ds_multiply
"""

Shared fields

See shared fields.

Unique fields

contractAddress: The Oracle or Operator contract to monitor for requests
requesters: Optional - Allows whitelisting requesters
minContractPaymentLinkJuels Optional - Allows you to specify a job-specific minimum contract payment
minIncomingConfirmations Optional - Allows you to specify a job-specific MIN_INCOMING_CONFIRMATIONS value, must be greater than global MIN_INCOMING_CONFIRMATIONS

Job type specific pipeline variables

$(jobSpec.databaseID): the ID of the job spec in the local database. You shouldn't need this in 99% of cases.
$(jobSpec.externalJobID): the globally-unique job ID for this job. Used to coordinate between node operators in certain cases.
$(jobSpec.name): the local name of the job.
$(jobRun.meta): a map of metadata that can be sent to a bridge, etc.
$(jobRun.logBlockHash): the block hash in which the initiating log was received.
$(jobRun.logBlockNumber): the block number in which the initiating log was received.
$(jobRun.logTxHash): the transaction hash that generated the initiating log.
$(jobRun.logAddress): the address of the contract to which the initiating transaction was sent.
$(jobRun.logTopics): the log's topics (indexed fields).
$(jobRun.logData): the log's data (non-indexed fields).
$(jobRun.blockReceiptsRoot) : the root of the receipts trie of the block (hash).
$(jobRun.blockTransactionsRoot) : the root of the transaction trie of the block (hash).
$(jobRun.blockStateRoot) : the root of the final state trie of the block (hash).

Examples

Get > Uint256 job

Let's assume that a user makes a request to an oracle to call a public API, retrieve a number from the response, remove any decimals and return uint256.

The smart contract example can be found here.
The job spec example can be found here.

Get > Int256 job

Let's assume that a user makes a request to an oracle to call a public API, retrieve a number from the response, remove any decimals and return int256.

The job spec example can be found here.

Get > Bool job

Let's assume that a user makes a request to an oracle to call a public API, retrieve a boolean from the response and return bool.

The job spec example can be found here.

Get > String job

Let's assume that a user makes a request to an oracle and would like to fetch a string from the response.

The smart contract example can be found here.
The job spec example can be found here.

Get > Bytes job

Let's assume that a user makes a request to an oracle and would like to fetch bytes from the response (meaning a response that contains an arbitrary-length raw byte data).

The smart contract example can be found here.
The job spec example can be found here.

Multi-Word job

Let's assume that a user makes a request to an oracle and would like to fetch multiple words in one single request.

The smart contract example can be found here.
The job spec example can be found here.

Existing job

Using an existing Oracle Job makes your smart contract code more succinct. Let's assume that a user makes a request to an oracle that leverages Etherscan External Adapter to retrieve the gas price.

The smart contract example can be found here.
The job spec example can be found here.

Flux Monitor Jobs

The Flux Monitor job type is for continually-updating data feeds that aggregate responses from multiple oracles. The oracles servicing the feed submit rounds based on several triggers:

An occasional poll, which must show that there has been sufficient deviation from an offchain data source before a new result is submitted
New rounds initiated by other oracles on the feeds. If another oracle notices sufficient deviation, all other oracles will submit their current observations as well.
A heartbeat, which ensures that even if no deviation occurs, we submit a new result to prove liveness. This can take one of two forms:
- The "idle timer", which begins counting down each time a round is started
- The "drumbeat", which simply ticks at a steady interval, much like a cron job

Spec format

type              = "fluxmonitor"
schemaVersion     = 1
name              = "example flux monitor spec"
contractAddress   = "0x3cCad4715152693fE3BC4460591e3D3Fbd071b42"
externalJobID     = "0EEC7E1D-D0D2-476C-A1A8-72DFB6633F03"

threshold = 0.5
absoluteThreshold = 0.0 # optional

idleTimerPeriod   = "1s"
idleTimerDisabled = false

pollTimerPeriod   = "1m"
pollTimerDisabled = false

drumbeatEnabled  = true
drumbeatSchedule = "CRON_TZ=UTC * */20 * * * *"

observationSource = """
    // data source 1
    ds1 [type="http" method=GET url="https://pricesource1.com"
         requestData="{\\"coin\\": \\"ETH\\", \\"market\\": \\"USD\\"}"]
    ds1_parse [type="jsonparse" path="data,result"]

    // data source 2
    ds2 [type="http" method=GET url="https://pricesource2.com"
         requestData="{\\"coin\\": \\"ETH\\", \\"market\\": \\"USD\\"}"]
    ds2_parse [type="jsonparse" path="data,result"]

    ds1 -> ds1_parse -> medianized_answer
    ds2 -> ds2_parse -> medianized_answer

    medianized_answer [type=median]
"""

Shared fields

See shared fields.

Unique fields

contractAddress: the address of the FluxAggregator contract that manages the feed.
threshold: the percentage threshold of deviation from the previous onchain answer that must be observed before a new set of observations are submitted to the contract.
absoluteThreshold: the absolute numerical deviation that must be observed from the previous onchain answer before a new set of observations are submitted to the contract. This is primarily useful with data that can legitimately sometimes hit 0, as it's impossible to calculate a percentage deviation from 0.
idleTimerPeriod: the amount of time (after the start of the last round) after which a new round will be automatically initiated, regardless of any observed offchain deviation.
idleTimerDisabled: whether the idle timer is used to trigger new rounds.
drumbeatEnabled: whether the drumbeat is used to trigger new rounds.
drumbeatSchedule: the cron schedule of the drumbeat. This field supports the same syntax as the cron job type (see the cron library documentation for details). CRON_TZ is required.
pollTimerPeriod: the frequency with which the offchain data source is checked for deviation against the previously submitted onchain answer.
pollTimerDisabled: whether the occasional deviation check is used to trigger new rounds.
Notes:
- For duration parameters, the maximum unit of time is h (hour). Durations of a day or longer must be expressed in hours.
- If no time unit is provided, the default unit is nanoseconds, which is almost never what you want.

Job type specific pipeline variables

$(jobSpec.databaseID): the ID of the job spec in the local database. You shouldn't need this in 99% of cases.
$(jobSpec.externalJobID): the globally-unique job ID for this job. Used to coordinate between node operators in certain cases.
$(jobSpec.name): the local name of the job.
$(jobRun.meta): a map of metadata that can be sent to a bridge, etc.

Keeper jobs

Keeper jobs occasionally poll a smart contract method that expresses whether something in the contract is ready for some onchain action to be performed. When it's ready, the job executes that onchain action.

Examples:

Liquidations
Rebalancing portfolios
Rebase token supply adjustments
Auto-compounding
Limit orders

Spec format

type            = "keeper"
schemaVersion   = 1
evmChainID      = 1
name            = "example keeper spec"
contractAddress = "0x7b3EC232b08BD7b4b3305BE0C044D907B2DF960B"
fromAddress     = "0xa8037A20989AFcBC51798de9762b351D63ff462e"

Shared fields

See shared fields.

Unique fields

evmChainID: The numeric chain ID of the chain on which Chainlink Automation Registry is deployed
contractAddress: The address of the Chainlink Automation Registry contract to poll and update
fromAddress: The Oracle node address from which to send updates
externalJobID: This is an optional field. When omitted it will be generated

Offchain reporting jobs

Offchain Reporting (OCR) jobs are used very similarly to Flux Monitor jobs. They update data feeds with aggregated data from many Chainlink oracle nodes. However, they do this aggregation using a cryptographically-secure offchain protocol that makes it possible for only a single node to submit all answers from all participating nodes during each round (with proofs that the other nodes' answers were legitimately provided by those nodes), which saves a significant amount of gas.

Offchain reporting jobs require the FEATURE_OFFCHAIN_REPORTING=true environment variable.

Bootstrap node

Every OCR cluster requires at least one bootstrap node as a kind of "rallying point" that enables the other nodes to find one another. Bootstrap nodes do not participate in the aggregation protocol and do not submit answers to the feed.

Spec format

type               = "offchainreporting"
schemaVersion      = 1
evmChainID         = 1
contractAddress    = "0x27548a32b9aD5D64c5945EaE9Da5337bc3169D15"
p2pBootstrapPeers  = [
    "/dns4/chain.link/tcp/1234/p2p/16Uiu2HAm58SP7UL8zsnpeuwHfytLocaqgnyaYKP8wu7qRdrixLju",
]
isBootstrapPeer = true
externalJobID   = "0EEC7E1D-D0D2-476C-A1A8-72DFB6633F05"

Shared fields

See shared fields.

Unique fields

contractAddress: The address of the OffchainReportingAggregator contract.
evmChainID: The chain ID of the EVM chain in which the job will operate.
p2pBootstrapPeers: A list of libp2p dial addresses of the other bootstrap nodes helping oracle nodes find one another on the network. It is used with P2P networking stack V1 as follows:

p2pBootstrapPeers = [ "/dns4/HOST_NAME_OR_IP/tcp/PORT/p2p/BOOTSTRAP_NODE'S_P2P_ID" ]
p2pv2Boostrappers: A list of libp2p dial addresses of the other bootstrap nodes helping oracle nodes find one another on the network. It is used with P2P networking stack V2 as follows:

p2pv2Boostrappers = [ "BOOTSTRAP_NODE'S_P2P_ID@HOST_NAME_OR_IP:PORT" ]
isBootstrapPeer: This must be set to true.

Job type specific pipeline variables

$(jobSpec.databaseID): The ID of the job spec in the local database. You shouldn't need this in 99% of cases.
$(jobSpec.externalJobID): The globally-unique job ID for this job. Used to coordinate between node operators in certain cases.
$(jobSpec.name): The local name of the job.
$(jobRun.meta): A map of metadata that can be sent to a bridge, etc.

Oracle node

Oracle nodes, on the other hand, are responsible for submitting answers.

Spec format

type               = "offchainreporting"
schemaVersion      = 1
evmChainID         = 1
name               = "OCR: ETH/USD"
contractAddress    = "0x613a38AC1659769640aaE063C651F48E0250454C"
externalJobID      = "0EEC7E1D-D0D2-476C-A1A8-72DFB6633F06"
p2pPeerID          = "12D3KooWApUJaQB2saFjyEUfq6BmysnsSnhLnY5CF9tURYVKgoXK"
p2pBootstrapPeers  = [
    "/dns4/chain.link/tcp/1234/p2p/16Uiu2HAm58SP7UL8zsnpeuwHfytLocaqgnyaYKP8wu7qRdrixLju",
]
isBootstrapPeer    = false
keyBundleID        = "7f993fb701b3410b1f6e8d4d93a7462754d24609b9b31a4fe64a0cb475a4d934"
monitoringEndpoint = "chain.link:4321"
transmitterAddress = "0xF67D0290337bca0847005C7ffD1BC75BA9AAE6e4"
observationTimeout = "10s"
blockchainTimeout  = "20s"
contractConfigTrackerSubscribeInterval = "2m"
contractConfigTrackerPollInterval = "1m"
contractConfigConfirmations = 3
observationSource = """
    // data source 1
    ds1          [type="bridge" name=eth_usd]
    ds1_parse    [type="jsonparse" path="one,two"]
    ds1_multiply [type="multiply" times=100]

    // data source 2
    ds2          [type="http" method=GET url="https://chain.link/eth_usd"
                  requestData="{\\"hi\\": \\"hello\\"}"]
    ds2_parse    [type="jsonparse" path="three,four"]
    ds2_multiply [type="multiply" times=100]

    ds1 -> ds1_parse -> ds1_multiply -> answer
    ds2 -> ds2_parse -> ds2_multiply -> answer

    answer [type=median]
"""

Shared fields

See shared fields.

Unique fields

contractAddress: The address of the OffchainReportingAggregator contract.
evmChainID: The chain ID of the EVM chain in which the job will operate.
p2pPeerID: The base58-encoded libp2p public key of this node.
p2pBootstrapPeers: A list of libp2p dial addresses of the other bootstrap nodes helping oracle nodes find one another on the network. It is used with P2P networking stack V1 as follows: p2pBootstrapPeers = [ "/dns4/<host name or ip>/tcp/<port>/p2p/<bootstrap node's P2P ID>" ]
p2pv2Boostrappers: A list of libp2p dial addresses of the other bootstrap nodes helping oracle nodes find one another on the network. It is used with P2P networking stack V2 as follows:

p2pv2Boostrappers = [ "<bootstrap node's P2P ID>@<host name or ip>:<port>" ]
keyBundleID: The hash of the OCR key bundle to be used by this node. The Chainlink node keystore manages these key bundles. Use the node Key Management UI or the chainlink keys ocr sub-commands in the CLI to create and manage key bundles.
monitoringEndpoint: The URL of the telemetry endpoint to send OCR metrics to.
transmitterAddress: The Ethereum address from which to send aggregated submissions to the OCR contract.
observationTimeout: The maximum duration to wait before an offchain request for data is considered to be failed/unfulfillable.
blockchainTimeout: The maximum duration to wait before an onchain request for data is considered to be failed/unfulfillable.
contractConfigTrackerSubscribeInterval: The interval at which to retry subscribing to onchain config changes if a subscription has not yet successfully been made.
contractConfigTrackerPollInterval: The interval at which to proactively poll the onchain config for changes.
contractConfigConfirmations: The number of blocks to wait after an onchain config change before considering it worthy of acting upon.

Job type specific pipeline variables

$(jobSpec.databaseID): The ID of the job spec in the local database. You shouldn't need this in 99% of cases.
$(jobSpec.externalJobID): The globally-unique job ID for this job. Used to coordinate between node operators in certain cases.
$(jobSpec.name): The local name of the job.
$(jobRun.meta): A map of metadata that can be sent to a bridge, etc.

Webhook Jobs

Webhook jobs can be initiated by HTTP request, either by a user or external initiator.

This is an example webhook job:

type            = "webhook"
schemaVersion   = 1
externalInitiators = [
  { name = "my-external-initiator-1", spec = "{\"foo\": 42}" },
  { name = "my-external-initiator-2", spec = "{}" }
]
observationSource   = """
    parse_request  [type="jsonparse" path="data,result" data="$(jobRun.requestBody)"]
    multiply       [type="multiply" input="$(parse_request)" times="100"]
    send_to_bridge [type="bridge" name="my_bridge" requestData="{ \\"result\\": $(multiply) }"]

    parse_request -> multiply -> send_to_bridge
"""

All webhook jobs can have runs triggered by a logged in user.

Webhook jobs may additionally specify zero or more external initiators, which can also trigger runs for this job. The name must exactly match the name of the referred external initiator. The external initiator definition here must contain a spec which defines the JSON payload that will be sent to the External Initiator on job creation if the external initiator has a URL. If you don't care about the spec, you can simply use the empty JSON object.

Unique fields

externalInitiators - an array of {name, spec} objects, where name is the name registered with the node, and spec is the job spec to be forwarded to the external initiator when it is created.

Shared fields

See shared fields.

Job type specific pipeline variables

$(jobSpec.databaseID): the ID of the job spec in the local database. You shouldn't need this in 99% of cases.
$(jobSpec.externalJobID): the globally-unique job ID for this job. Used to coordinate between node operators in certain cases.
$(jobSpec.name): the local name of the job.
$(jobRun.meta): a map of metadata that can be sent to a bridge, etc.
$(jobRun.requestBody): the body of the request that initiated the job run.

Job Types

Get the latest Chainlink content straight to your inbox.