Loading

Supported technologies

The Elastic APM Python Agent supports the technologies listed below.

Elastic supports OpenTelemetry, which allows logs, metrics, and trace signal collection for many of these technologies. Consider using the EDOT Python SDK for observability data so you continue to get the full power of Elastic's platform.

The Elastic APM Python Agent comes with support for the following frameworks:

For other frameworks and custom Python code, the agent exposes a set of APIs for integration.

The following Python versions are supported:

  • 3.6
  • 3.7
  • 3.8
  • 3.9
  • 3.10
  • 3.11
  • 3.12
  • 3.13

We support these Django versions:

  • 1.11
  • 2.0
  • 2.1
  • 2.2
  • 3.0
  • 3.1
  • 3.2
  • 4.0
  • 4.2
  • 5.0

For upcoming Django versions, we generally aim to ensure compatibility starting with the first Release Candidate.

Note

we currently don’t support Django running in ASGI mode.

We support these Flask versions:

  • 0.10 (Deprecated)
  • 0.11 (Deprecated)
  • 0.12 (Deprecated)
  • 1.0
  • 1.1
  • 2.0
  • 2.1
  • 2.2
  • 2.3
  • 3.0

We support these aiohttp versions:

  • 3.x

We support these tornado versions:

  • 6.x

We support these sanic versions:

  • 20.12.2,<26

We support these Starlette versions:

  • 0.13.0,<1

Any FastAPI version which uses a supported Starlette version should also be supported.

We support these grpcio versions:

  • 1.24.0,<2

The Python APM agent comes with automatic instrumentation of various 3rd party modules and standard library modules.

We support these Celery versions:

  • 4.x (deprecated)
  • 5.x

Celery tasks will be recorded automatically with Django and Flask only.

Instrumented methods:

  • elasticsearch.transport.Transport.perform_request
  • elasticsearch.connection.http_urllib3.Urllib3HttpConnection.perform_request
  • elasticsearch.connection.http_requests.RequestsHttpConnection.perform_request
  • elasticsearch._async.transport.AsyncTransport.perform_request
  • elasticsearch_async.connection.AIOHttpConnection.perform_request

Additionally, the instrumentation wraps the following methods of the Elasticsearch client class:

  • elasticsearch.client.Elasticsearch.delete_by_query
  • elasticsearch.client.Elasticsearch.search
  • elasticsearch.client.Elasticsearch.count
  • elasticsearch.client.Elasticsearch.update

Collected trace data:

  • the query string (if available)
  • the query element from the request body (if available)
  • the response status code
  • the count of affected rows (if available)

We recommend using keyword arguments only with elasticsearch-py, as recommended by the elasticsearch-py docs. If you are using positional arguments, we will be unable to gather the query element from the request body.

Instrumented methods:

  • sqlite3.connect
  • sqlite3.dbapi2.connect
  • pysqlite2.dbapi2.connect

The instrumented connect method returns a wrapped connection/cursor which instruments the actual Cursor.execute calls.

Collected trace data:

  • parametrized SQL query

Library: MySQLdb (<2)

Instrumented methods:

  • MySQLdb.connect

The instrumented connect method returns a wrapped connection/cursor which instruments the actual Cursor.execute calls.

Collected trace data:

  • parametrized SQL query

Library: mysql-connector-python (<9)

Instrumented methods:

  • mysql.connector.connect

The instrumented connect method returns a wrapped connection/cursor which instruments the actual Cursor.execute calls.

Collected trace data:

  • parametrized SQL query

Library: pymysql (<2)

Instrumented methods:

  • pymysql.connect

The instrumented connect method returns a wrapped connection/cursor which instruments the actual Cursor.execute calls.

Collected trace data:

  • parametrized SQL query

Library: aiomysql (<1)

Instrumented methods:

  • aiomysql.cursors.Cursor.execute

Collected trace data:

  • parametrized SQL query

Library: psycopg2, psycopg2-binary (>=2.9,<3)

Instrumented methods:

  • psycopg2.connect

The instrumented connect method returns a wrapped connection/cursor which instruments the actual Cursor.execute calls.

Collected trace data:

  • parametrized SQL query

Library: psycopg, psycopg-binary (>3.0.0,<4)

Instrumented methods:

  • psycopg.connect

The instrumented connect method returns a wrapped connection/cursor which instruments the actual Cursor.execute calls.

Collected trace data:

  • parametrized SQL query

Library: aiopg (>=1.0,<2)

Instrumented methods:

  • aiopg.cursor.Cursor.execute
  • aiopg.cursor.Cursor.callproc

Collected trace data:

  • parametrized SQL query

Library: asyncpg (>=0.20,<2)

Instrumented methods:

  • asyncpg.connection.Connection.execute
  • asyncpg.connection.Connection.executemany

Collected trace data:

  • parametrized SQL query

Library: pyodbc (>=4.0,<6)

Instrumented methods:

  • pyodbc.connect

The instrumented connect method returns a wrapped connection/cursor which instruments the actual Cursor.execute calls.

Collected trace data:

  • parametrized SQL query

Library: pymssql (>=2.1.0,<3)

Instrumented methods:

  • pymssql.connect

The instrumented connect method returns a wrapped connection/cursor which instruments the actual Cursor.execute calls.

Collected trace data:

  • parametrized SQL query

Library: pymongo (>=2.9,<5)

Instrumented methods:

  • pymongo.collection.Collection.aggregate
  • pymongo.collection.Collection.bulk_write
  • pymongo.collection.Collection.count
  • pymongo.collection.Collection.create_index
  • pymongo.collection.Collection.create_indexes
  • pymongo.collection.Collection.delete_many
  • pymongo.collection.Collection.delete_one
  • pymongo.collection.Collection.distinct
  • pymongo.collection.Collection.drop
  • pymongo.collection.Collection.drop_index
  • pymongo.collection.Collection.drop_indexes
  • pymongo.collection.Collection.ensure_index
  • pymongo.collection.Collection.find_and_modify
  • pymongo.collection.Collection.find_one
  • pymongo.collection.Collection.find_one_and_delete
  • pymongo.collection.Collection.find_one_and_replace
  • pymongo.collection.Collection.find_one_and_update
  • pymongo.collection.Collection.group
  • pymongo.collection.Collection.inline_map_reduce
  • pymongo.collection.Collection.insert
  • pymongo.collection.Collection.insert_many
  • pymongo.collection.Collection.insert_one
  • pymongo.collection.Collection.map_reduce
  • pymongo.collection.Collection.reindex
  • pymongo.collection.Collection.remove
  • pymongo.collection.Collection.rename
  • pymongo.collection.Collection.replace_one
  • pymongo.collection.Collection.save
  • pymongo.collection.Collection.update
  • pymongo.collection.Collection.update_many
  • pymongo.collection.Collection.update_one

Collected trace data:

  • database name
  • method name

Library: redis (>=2.8,<8)

Instrumented methods:

  • redis.client.Redis.execute_command
  • redis.client.Pipeline.execute

Collected trace data:

  • Redis command name

Library: aioredis (<=2.0.1)

Instrumented methods:

  • aioredis.pool.ConnectionsPool.execute
  • aioredis.commands.transaction.Pipeline.execute
  • aioredis.connection.RedisConnection.execute

Collected trace data:

  • Redis command name

Library: cassandra-driver (>=3.24,<4.0)

Instrumented methods:

  • cassandra.cluster.Session.execute
  • cassandra.cluster.Cluster.connect

Collected trace data:

  • CQL query

Library: python-memcached (>=1.51,<2)

Instrumented methods:

  • memcache.Client.add
  • memcache.Client.append
  • memcache.Client.cas
  • memcache.Client.decr
  • memcache.Client.delete
  • memcache.Client.delete_multi
  • memcache.Client.disconnect_all
  • memcache.Client.flush_all
  • memcache.Client.get
  • memcache.Client.get_multi
  • memcache.Client.get_slabs
  • memcache.Client.get_stats
  • memcache.Client.gets
  • memcache.Client.incr
  • memcache.Client.prepend
  • memcache.Client.replace
  • memcache.Client.set
  • memcache.Client.set_multi
  • memcache.Client.touch

Collected trace data:

  • Destination (address and port)

Library: pymemcache (>=3.0,<4.1)

Instrumented methods:

  • pymemcache.client.base.Client.add
  • pymemcache.client.base.Client.append
  • pymemcache.client.base.Client.cas
  • pymemcache.client.base.Client.decr
  • pymemcache.client.base.Client.delete
  • pymemcache.client.base.Client.delete_many
  • pymemcache.client.base.Client.delete_multi
  • pymemcache.client.base.Client.flush_all
  • pymemcache.client.base.Client.get
  • pymemcache.client.base.Client.get_many
  • pymemcache.client.base.Client.get_multi
  • pymemcache.client.base.Client.gets
  • pymemcache.client.base.Client.gets_many
  • pymemcache.client.base.Client.incr
  • pymemcache.client.base.Client.prepend
  • pymemcache.client.base.Client.quit
  • pymemcache.client.base.Client.replace
  • pymemcache.client.base.Client.set
  • pymemcache.client.base.Client.set_many
  • pymemcache.client.base.Client.set_multi
  • pymemcache.client.base.Client.stats
  • pymemcache.client.base.Client.touch

Collected trace data:

  • Destination (address and port)

Library: kafka-python (>=2.0,<3)

Instrumented methods:

  • kafka.KafkaProducer.send,
  • kafka.KafkaConsumer.poll,
  • kafka.KafkaConsumer.__next__

Collected trace data:

  • Destination (address and port)
  • topic (if applicable)

Library: urllib.request (Python 3)

Instrumented methods:

  • urllib.request.AbstractHTTPHandler.do_open

Collected trace data:

  • HTTP method
  • requested URL

Library: urllib3 (<3)

Instrumented methods:

  • urllib3.connectionpool.HTTPConnectionPool.urlopen

Additionally, we instrumented vendored instances of urllib3 in the following libraries:

  • requests
  • botocore

Both libraries have "unvendored" urllib3 in more recent versions, we recommend to use the newest versions.

Collected trace data:

  • HTTP method
  • requested URL

Library: requests (<3)

Instrumented methods:

  • requests.sessions.Session.send

Collected trace data:

  • HTTP method
  • requested URL

Library: aiohttp (>=3,<4)

Instrumented methods:

  • aiohttp.client.ClientSession._request

Collected trace data:

  • HTTP method
  • requested URL

Library: httpx (<1)

Instrumented methods:

  • `httpx.Client.send

Collected trace data:

  • HTTP method
  • requested URL

Library: boto3 (>=1.0,<2)

Instrumented methods:

  • botocore.client.BaseClient._make_api_call

Collected trace data for all services:

  • AWS region (e.g. eu-central-1)
  • AWS service name (e.g. s3)
  • operation name (e.g. ListBuckets)

Additionally, some services collect more specific data

Library: aiobotocore (>=2.2.0,<3)

Instrumented methods:

  • aiobotocore.client.BaseClient._make_api_call

Collected trace data for all services:

  • AWS region (e.g. eu-central-1)
  • AWS service name (e.g. s3)
  • operation name (e.g. ListBuckets)

Additionally, some services collect more specific data

  • Bucket name
  • Table name
  • Topic name
  • Queue name

Library: Django (see Django for supported versions)

Instrumented methods:

  • django.template.Template.render

Collected trace data:

  • template name

Library: jinja2

Instrumented methods:

  • jinja2.Template.render

Collected trace data:

  • template name