Reading Logs

Each record in the service logs has the form:

LEVEL <request_guid> [timestamp] message

Example:

INFO <57d07e28-78b6-428a-995a-88e8d7c67f2e> [2020-01-01 00:05:08,147] com.tamr.dataset.client.DatasetClient: Querying provider http://<ip_address>:9150/api for datasets

Each record in the request logs has the form:

<request guid> code - - [timestamp] “<HTTP Method> <URL>” <http response status code> <time to process in ms>

Example:

<dbf2fbe5-8f0a-4d99-a1db-5acef9bd2544> 10.30.0.12 - - [01/Jan/2020:18:19:22 +0000] "POST /api/dataset/datasets/query HTTP/1.1" 200 5212 "http://<ip_address>:9100/" "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.97 Safari/537.36" 26

You can track related information across different logs using the request guid, for example:

<tamr_home_directory>/tamr/logs$ grep "57d07e28-78b6-428a-995a-88e8d7c67f2e" ./*

Service and Spark logs can be configured into JSON format by setting TAMR_LOG_JSON_ENABLED to True. This may make logs easier to index with log search tools.

Navigating Logs

Tamr Core and its microservices produce many different logs. These are some of the most common logs to start with when troubleshooting:

• unify.log: Often worth checking first. Backup and restore information and errors related to the startup of Tamr Core and its dependencies.
• auth.log: Auth service logs. Error information related to authentication/authorization operations.
• dataset.log: Dataset service logs. Error information related to dataset operations, such as creating datasets, inserting/updating records in source datasets, and running jobs to update Unified Datasets.
• recipe.log: Recipe logs. Error information related to recipe operations.
• dedup.log: Dedup service logs. Error information related to dedup operations.

If you are unsure which log to start with, you can check the last modified log files under <tamr_home_directory>/tamr/logs. The linux command ls -ltr will display all of the files in your current directory and will sort the response so most recent files are at the bottom.

You can also use grep to find files containing ERROR level logs and error messages, for example grep -i “error” *.log, or check whether a specific log contains an error: grep -i “error” unify.log.

Kibana

Kibana is part of the Tamr Core monitoring stack and allows you to see some log information. When troubleshooting you can search for the word "error" to see all corresponding log entries. See Monitoring.

See Also

• Logging in Single-Node On-Premises Deployments
• Logging in Cloud Platform Deployments