Extraction logs provide detail about the extraction portion of the replication process for a given integration. In this guide, we’ll cover:

Log retention

Extraction logs are grouped by day. The number of days’ worth of logs available to you depends on your Stitch plan:

  • Free trial: 1 day

  • Standard: 7 days

  • Advanced: 60 days

  • Premium: 60 days

Logs and plan changes

Changing your plan can impact logs currently available to you.

Plan downgrades

If you downgrade to a plan that offers fewer days’ logs, you’ll lose access to the difference between your current plan and your new plan.

For example: If you downgrade to Standard from the Advanced plan, you’ll lose access to 53 days’ worth of logs.

Plan upgrades

Likewise, if you upgrade to a plan that offers more days’ logs, you’ll immediately gain access to the difference.

For example: If you upgrade to Advanced from the Standard plan, you’ll gain access to an additional 53 days’ worth of logs.

Log composition

The graph at the top of the Extractions tab displays every time Stitch connected to the integration by day, based on the integration’s Replication Frequency.

Extraction graph with intervals of 30 minutes

In addition to displaying the time an extraction began, the tooltips also include how long the extraction ran for and if any errors arose.

To view the raw logs for a specific extraction, click the View Logs link in the tooltip or the bar in the graph. This will open the logs for the job in the Extraction Logs section, located below the graph.

Logs for in progress jobs

When a replication job is currently running, the job will have an In Progress status and a Stop Extraction button:

An In Progress replication job and Stop Extraction button, highlighted

You can use the Stop Extraction button to stop an in progress job, which is useful when the data source is experiencing issues. Refer to the Start and stop extraction jobs documentation for more info.

Extraction log fields

Lines in raw extraction logs are made up of four fields. Before we get into the field details, take a look at this example line:

2017-11-17 16:44:41,159Z tap - INFO State update: adding bookmarks.ads.updated_time = "2017-11-06T13:29:23-05:00"

In the table below, we’ll break down each of the fields and explain what they mean.

Order Name Example Description
1 Timestamp

2017-11-17 16:44:41,159Z

The date and time the action occurred.

2 Location


Indicates the part of the integration the message comes from. This is used only by Stitch support to debug issues.

Possible values are:

  • main
  • tap
  • target
3 Message type


The message type. In general:

  • INFO indicates things are working as expected
  • DEBUG indicates info of interest when diagnosing issues
  • WARNING indicates something unexpected happened, but extraction continued. This status isn’t usually noteworthy unless a CRITICAL error is also returned.
  • ERROR indicates the integration was unable to perform some function, but extraction continued. This status isn’t usually noteworthy unless a CRITICAL error is also returned.
  • CRITICAL indicates an error occurred that caused the extraction to terminate. See Extraction Errors for more info.
4 Message body

State update: adding bookmarks.ads.updated_time = "2017-11-06T13:29:23-05:00"

A description of the action currently being taken.

Extraction log examples

When reading the extraction logs for your integrations, pay particular attention to the content of the message body. The message body will contain information about what’s currently happening in the extraction process and errors, should they arise.

Below are some examples of extraction logs, what they indicate, and how to read them.

Replication Key values

Replication Keys are columns used to identify new and updated data in tables that use Key-based Incremental Replication.

The extraction logs contain information about the current Replication Key value for a given table, as well as the updated value detected during the extraction process.

Replication Key values and extraction

The last saved maximum Replication Key value for a given table is used to detect new and updated data.

For example: In this log line, Stitch will extract data from the ads endpoint that has an updated_at timestamp greater than or equal to 2017-11-06T12:48:15-05:00:

2017-11-21 18:16:08,389Z    tap - INFO found current bookmark for ads:  2017-11-06T12:48:15-05:00

Updated Replication Key values

During the extraction process, Stitch will update the Replication Key values for the tables set to replicate.

Generally, the log line containing the saved Replication Key value for a table will be similar to this:

2017-11-21 18:21:15,307Z   main - INFO State update: adding bookmarks.ads.updated_time = "2017-11-06T13:29:23-05:00"

Beginning after bookmarks, read the following as table_name.replication_key_field. In this case, the table is ads, the Replication Key is updated_time, and the saved Replication Key value for the table is being updated to 2017-11-06T13:29:23-05:00.

Note: This line will not display in the logs for tables that have unchanged Replication Key values. This means if Stitch doesn’t detect any new or updated data for a table, a line like the above will not appear in the logs for that table.

Extraction errors

If an error occurs that terminates the extraction process, a line with a message type of CRITICAL will appear in the log. Generally, this will also display as the last line of the log.

For example: During this extraction of Salesforce data, Stitch detected that there wasn’t sufficient API quota available to continue replication:

2017-11-20 07:48:45,410Z    tap - CRITICAL Salesforce has reported 32115/100000 (32.12%) total REST quota used across all Salesforce Applications. Terminating replication to not continue past configured percentage of 30.0% total quota.

For the majority of errors, Stitch will parse out and display the messages separately from the raw extraction logs:

Error message from error that occurred during extraction

If an error arises, check out the extraction error references for help.

Questions? Feedback?

Did this article help? If you have questions or feedback, feel free to submit a pull request with your suggestions, open an issue on GitHub, or reach out to us.