Pricing Login
Interactive demos

Click through interactive platform demos now.

Live demo, real expert

Schedule a platform demo with a Sumo Logic expert.

Start free trial
Back to blog results

April 29, 2020 By Rick Jury

“Fiel-ding Good” - Three great ways to enrich AWS logs in Sumo Logic

In this article I’m going to show you three quick and easy ways to enrich your AWS log data in Sumo Logic using fields:

  1. Report instanceid with each log event for EC2;
  2. Bring in AWS tags for EC2 instance logs
  3. Post custom tags for Cloudtrail events via AWS Lambda.

Fields is a feature of Sumo shipped in 2019 as part of our Kubernetes monitoring solution and is how the fluentd pipeline adds Kubernetes metadata like service and pod to each log event.

It’s also really handy for enriching each log event from AWS with fields such as:

  • Instance-id, region and AZ;
  • Auto scale group, application or build version;
  • Team, owner or environment.

Enriching your logs with these extra fields at ingest time will:

  • Make searching easier: because business attributes such as application name or owning team can be used in searches;
  • Make searches run faster: because searches can be easily targeted by built in fields, without needing to parse them at search time;
  • Make it easy to correlate logs with other tools using key attributes like instanceid or tags.

This translates into better business outcomes for you:

  • Improved observability of your AWS environment;
  • Faster times to detect and resolve issues;
  • New features built and shipped faster.

Nothing ever lasts forever: EC2 instance-ids

Did you know in less than 5 minutes you can configure your Sumo Logic collectors to report instanceid as a field for all log events?

When you install Sumo Logic installed collectors on EC2 instances, by default the collector name (_collector) and sourcehost (_sourcehost) default to the local host name. Unlike on-prem environments where hosts typically had meaningful host names, in AWS meaningless ephemeral host names are common. Often it’s hard to correlate these names with key attributes like instance-id, auto scaling group, or service name to troubleshoot issues. 

You can write a script to custom configure the collector name property in using instance metadata during instance start but there is an easier way … 

Since mid 2019 collectors installed on AWS EC2 agents report four fields by default to Sumo Logic: availabilityZone, instanceId, instanceType, and region. The fields just don’t appear unless you enable them.

If you have installed collectors on EC2 instances take a couple of minutes to enable the fields in Manage Settings / Fields in Sumo Logic UI. Open the Fields settings in Sumo UI, select ‘Dropped Fields’ and enable each of the four fields. It’s that simple!

New logs ingested from EC2 instances will have these four fields assigned. You can then search and group by these four fields in your searches as shown below:

A little souvenir: add AWS tags to your logs

If you are building in AWS it makes sense to invest in creating a best practice tagging strategy. As your AWS footprint grows these tags on EC2 instances will deliver increasing value: enabling you to categorize resources by costcenter, purpose, owner, environment, or other criteria. How wonderful would it be to search logs in Sumo Logic using your AWS tags you worked so hard to create?

A common approach is to encode values like application names, environments, or tag values into builtin sumo metadata fields like the collector name in, or configuring source category string in local json files for each source.

In 2019 Sumo Logic released a new source that provides a better approach for EC2- where you can map tags directly to ingeste logs. Configure an AWS Log Metadata Source and the EC2 instance AWS tags are mapped as fields to each log event at ingestion time.

The steps to setup the AWS metadata source can be found in the docs page - but here is a quick overview of the process:

  1. Decide on a list of tags you want to use in sumo. Say for example: name,  owner and costcenter
  2. Open the Manage Settings / Fields UI page in sumo and create a field for each key name you plan to use.
  3. Create an IAM role and policy with necessary permissions. When you create the AWS metadata source there is a button to generate a cloudformation template.
  4. Finish setting up the source in the Manage/Collection screen in Sumo Logic.

Shortly after setting up the log metadata source you will see tags appear as fields in newly ingested logs from EC2 instances in that AWS account. These fields will appear in the field browser on the left hand side of the sumo search UI.

You can then use these in searches for example:

_sourcecategory=aws/ec2/logs owner=myteam

You can also use these fields for aggregation and grouping:

_sourcecategory=aws/ec2/logs owner=myteam | sum(_size) by owner, cost_center

Scaling To Many Accounts

If you have 10s or 100s of AWS accounts you can create these AWS sources using a code pipeline via the sources management API. Here is an example JSON for a AWS metadata source.

{ "api.version":"v1", "source":{   "name":"aws_metadata_046921848075",   "description":"Poll metadata for acme-xxx-test 0123456789",   "automaticDateParsing":false,   "multilineProcessingEnabled":false,   "useAutolineMatching":false,   "contentType":"AwsMetadata",   "forceTimeZone":false,   "filters":[],   "cutoffTimestamp":0,   "encoding":"UTF-8",   "fields":{       },   "thirdPartyRef":{     "resources":[{       "serviceType":"AwsMetadata",       "path":{         "type":"AwsMetadataPath",         "limitToRegions":["us-east-1","us-west-2"],         "limitToNamespaces":["AWS/EC2"],         "tagFilters":["owner","costcenter","cluster","Name"]       },       "authentication":{         "type":"AWSRoleBasedAuthentication",         "roleARN":"arn:aws:iam::0123456789:role/sumologic-metadata-tags"       }     }]   },   "scanInterval":60000,   "paused":false,   "sourceType":"Polling" }}

I know you love to make an entrance: What about Cloudtrail?

Sumo fields can be posted to an HTTPS log source with custom headers to modify metadata including the X-Sumo-Fields header. X-Sumo-Fields is a comma separated list of key value pairs and we can leverage this feature to add tags or other fields to Cloudwatch logs.

Upload it: X-Sumo-Fields Demo

Let’s do a really quick demo of using fields in sumo - but feel free to skip this if you want to go straight to the Cloudwatch Lambda section!

Check field names exist and create if required

Each field you want to send must be defined in the fields setting screen. Go to Manage Data / Settings / Fields and create a new field name - for example: env,team, owner and version.

Create an HTTP source for testing POST

Create a hosted collector by going to Manage Data / Collection.

Then create a new HTTPS streaming source.

Take a note of the url to post data to this new source. You will need this in the next step.

POST log events data with metadata fields

Configure uri from the previous step as the endpoint in this script, save and execute to post data. For example:

export endpoint="https://my_url"./
# MAKE SURE TO SET A VALID ENDPOINT VARIABLE!endpoint="${endpoint:-}" # some random test data to make it interestingepoch=`date +%s`dt=`date`h=`hostname`deets=`uname -a` # defaults if we don't set with env varscategory="${category:-test/app/test123}"payload="${payload:-{\"timestamp\":$epoch,\"time_string\":\"$dt\",\"from_host\":\"$h\",\"deets\":\"$deets\",\"random_id\":\"$((1 + RANDOM % 100000))\"}}" # log metadata beta X-Sumo-Fields headerfields="${fields:-url=env=test,team=myteam,,version=0.0.1}" # this would send a payload Category (_sourcecategory) and Fields set in headers.curl -H "X-Sumo-Category:$category" -H "X-Sumo-Fields:$fields" -d "$payload" -X POST $endpoint

If you have a livetail session open you will see the events arrive immediately:

After a few minutes you can find your logs in a new search: 


In the UI you should see your new fields associated with your log on the left side of the screen in the field browser.  

We can now use our fields in any search scope or aggregation:

_sourcecategory=test/app/test123 team=* version=0.0.1 | count by team,owner,version

Using Fields With Cloudwatch Log Lambdas

Now you’ve seen how easy it is to post custom field data to Sumo consider how you could enhance your posting process for Cloudwatch logs via lambda or other custom log sources.

The Sumo Logic cloudwatch logs lambda function supports setting X-Sumo-Fields and other headers via an environment variable. This search shows the relevant code in the github repo.

Set the SOURCE_FIELDS_OVERRIDE env variable when the lambda executes to include your own custom field values.  

You can easily extend your lambda code to post custom fields to Sumo Logic for other use cases. For example creating code to read tags from AWS objects and include these as extra fields in each log event. It’s possible to auto subscribe lambda log groups and send your key tags as fields along with each event.

It’s a new world...building more value

I hope you enjoyed this brief tour of enriching your AWS logs for three use cases in Sumo Logic using Fields. If you take a little time to enrich your logs you will find you can search faster and deliver more value and observability with Sumo Logic.

Before you go: here are two ways you can leverage your new fields to build even more value:

  • Build templatized search templates to make it easy to search by key fields you created
  • Create dashboards with filters per application  / service / region so a single templated dashboard can be used for each new service you deploy. These could be launched from directly passing in the parameters in the url query string from other apps in your environment:

If you need a little bit more guidance, or have questions or feedback please reach out to

Complete visibility for DevSecOps

Reduce downtime and move from reactive to proactive monitoring.

Sumo Logic cloud-native SaaS analytics

Build, run, and secure modern applications and cloud infrastructures.

Start free trial
Rick Jury

Rick Jury

Customer Success Engineer

More posts by Rick Jury.

People who read this also enjoyed