Log Forwarding log-forwarding
Customers with a license with a logging vendor or who host a logging product can have AEM logs (including Apache/Dispatcher) and CDN logs forwarded to the associated logging destination. AEM as a Cloud Service supports the following logging destinations:
- Azure Blob Storage
- Datadog
- Elasticsearch or OpenSearch
- HTTPS
- Splunk
Log forwarding is configured in a self-service manner by declaring a configuration in Git, and deploying it via the Cloud Manager Config Pipeline to RDE, dev, stage, and production environment types in production (non-sandbox) programs.
There is an option for the AEM and Apache/Dispatcher logs to be routed through AEM’s advanced networking infrastructure, such as dedicated egress IP.
Note that the network bandwidth associated with logs sent to the logging destination are considered part of your organization’s Network I/O usage.
How This Article is Organized how-organized
This article is organized in the following way:
- Setup - common for all logging destinations
- Logging destination configurations - each destination has a slightly different format
- Log Entry Formats - information about the log entry formats
- Advanced networking - sending AEM and Apache/Dispatcher logs through a dedicated egress or through a VPN
- Migrating from legacy log forwarding - how to move from log forwarding previously setup by Adobe to the self-serve approach
Setup setup
-
Create a file named
logForwarding.yaml
. It should contain metadata, as described in the config pipeline article (kind should be set toLogForwarding
and version set to “1”), with a configuration similar to the following (we use Splunk as an example).code language-none kind: "LogForwarding" version: "1" metadata: envTypes: ["dev"] data: splunk: default: enabled: true host: "splunk-host.example.com" token: "${{SPLUNK_TOKEN}}" index: "AEMaaCS"
-
Place the file somewhere under a top level folder named config or similar, as described in Using Config Pipelines.
-
For environment types other than RDE (which uses command line tooling), create a targeted deployment config pipeline in Cloud Manager, as referenced by this section; note that Full Stack pipelines and Web Tier pipelines do not deploy the configuration file.
-
Deploy the configuration.
Tokens in the configuration (such as ${{SPLUNK_TOKEN}}
) represent secrets, which should not be stored in Git. Instead, declare them as Cloud Manager Secret Environment Variables. Make sure to select All as the dropdown value for the Service Applied field, so logs can be forwarded to author, publish, and preview tiers.
It is possible to set different values between CDN logs and AEM logs (including Apache/Dispatcher), by including an additional cdn and/or aem block after the default block, where properties can override those defined in the default block; only the enabled property is required. A possible use case could be to use a different Splunk index for CDN logs, as the example below illustrates.
kind: "LogForwarding"
version: "1"
metadata:
envTypes: ["dev"]
data:
splunk:
default:
enabled: true
host: "splunk-host.example.com"
token: "${{SPLUNK_TOKEN}}"
index: "AEMaaCS"
cdn:
enabled: true
token: "${{SPLUNK_TOKEN_CDN}}"
index: "AEMaaCS_CDN"
Another scenario is to disable either forwarding of the CDN logs or AEM logs (including Apache/Dispatcher). For example, to only forward the CDN logs, one can configure the following:
kind: "LogForwarding"
version: "1"
metadata:
envTypes: ["dev"]
data:
splunk:
default:
enabled: true
host: "splunk-host.example.com"
token: "${{SPLUNK_TOKEN}}"
index: "AEMaaCS"
aem:
enabled: false
Logging Destination Configuration logging-destinations
Configurations for the supported logging destinations are listed below, along with any specific considerations.
Azure Blob Storage azureblob
kind: "LogForwarding"
version: "1"
metadata:
envTypes: ["dev"]
data:
azureBlob:
default:
enabled: true
storageAccountName: "example_acc"
container: "aem_logs"
sasToken: "${{AZURE_BLOB_SAS_TOKEN}}
A SAS token should be used for authentication. It should be created from the Shared access signature page, rather than on the Shared access token page, and should be configured with these settings:
- Allowed services: Blob must be selected.
- Allowed resources: Object must be selected.
- Allowed permissions: Write, Add, Create must be selected.
- A valid Start and Expiry date/time.
Here is a screenshot of a sample SAS token configuration:
If logs have stopped being delivered after previously functioning correctly, check whether the SAS token you configured is still valid, as it may have expired.
Azure Blob Storage CDN logs azureblob-cdn
Each of the globally distributed logging servers will produce a new file every few seconds, under the aemcdn
folder. Once created, the file will no longer be appended to. The filename format is YYYY-MM-DDThhss.sss-uniqueid.log. E.g., 2024-03-04T10:00:00.000-WnFWYN9BpOUs2aOVn4ee.log.
As an example, at some point in time:
aemcdn/
2024-03-04T10:00:00.000-abc.log
2024-03-04T10:00:00.000-def.log
And then 30 seconds later:
aemcdn/
2024-03-04T10:00:00.000-abc.log
2024-03-04T10:00:00.000-def.log
2024-03-04T10:00:30.000-ghi.log
2024-03-04T10:00:30.000-jkl.log
2024-03-04T10:00:30.000-mno.log
Each file contains multiple json log entries, each on a separate line. The log entry formats are described under Logging for AEM as a Cloud Service, and each log entry also includes the additional properties mentioned in the Log Entry Formats section below.
Azure Blob Storage AEM logs azureblob-aem
AEM logs (including Apache/Dispatcher) appear below a folder with the following naming convention:
- aemaccess
- aemerror
- aemrequest
- aemdispatcher
- aemhttpdaccess
- aemhttpderror
Under each folder, a single file will be created and appended to. Customers are responsible for processing and managing this file so it does not grow too large.
See the log entry formats under Logging for AEM as a Cloud Service. The log entries will also include the additional properties mentioned in the Log Entry Formats section below.
Datadog datadog
kind: "LogForwarding"
version: "1"
metadata:
envTypes: ["dev"]
data:
datadog:
default:
enabled: true
host: "http-intake.logs.datadoghq.eu"
token: "${{DATADOG_API_KEY}}"
tags:
tag1: value1
tag2: value2
Considerations:
- Create an API Key, without any integration with a specific cloud provider.
- The tags property is optional
- For AEM logs, the Datadog source tag is set to one of
aemaccess
,aemerror
,aemrequest
,aemdispatcher
,aemhttpdaccess
, oraemhttpderror
- For CDN logs, the Datadog source tag is set to
aemcdn
- The Datadog service tag is set to
adobeaemcloud
, but you can overwrite it in the tags section - If your ingestion pipeline uses Datadog tags to determine the appropriate index for forwarding logs, verify that these tags are correctly configured in the Log Forwarding YAML file. Missing tags may prevent successful log ingestion if the pipeline depends on them.
Elasticsearch and OpenSearch elastic
kind: "LogForwarding"
version: "1"
metadata:
envTypes: ["dev"]
data:
elasticsearch:
default:
enabled: true
host: "example.com"
user: "${{ELASTICSEARCH_USER}}"
password: "${{ELASTICSEARCH_PASSWORD}}"
pipeline: "ingest pipeline name"
Considerations:
- by default, the port is 443. It can optionally be overridden with a property named
port
- For credentials, make sure to use use deployment credentials, rather than account credentials. These are the credentials that are generated in a screen that may resemble this image:
- For AEM logs,
index
is set to one ofaemaccess
,aemerror
,aemrequest
,aemdispatcher
,aemhttpdaccess
, oraemhttpderror
- The optional pipeline property should be set to the name of the Elasticsearch or OpenSearch ingest pipeline, which can be configured to route the log entry to the appropriate index. The pipeline’s Processor type must be set to script and the script language should be set to painless. Here is a sample script snippet to route log entries into an index such as aemaccess_dev_26_06_2024:
def envType = ctx.aem_env_type != null ? ctx.aem_env_type : 'unknown';
def sourceType = ctx._index;
def date = new SimpleDateFormat('dd_MM_yyyy').format(new Date());
ctx._index = sourceType + "_" + envType + "_" + date;
HTTPS https
kind: "LogForwarding"
version: "1"
metadata:
envTypes: ["dev"]
data:
https:
default:
enabled: true
url: "https://example.com/aem_logs/aem"
authHeaderName: "X-AEMaaCS-Log-Forwarding-Token"
authHeaderValue: "${{HTTPS_LOG_FORWARDING_TOKEN}}"
Considerations:
- The url string must include https:// or validation will fail.
- The url may include a port. For example,
https://example.com:8443/aem_logs/aem
. If no port is included in the url string, port 443 (the default HTTPS port) is assumed.
HTTPS CDN logs https-cdn
Web requests (POSTs) will be sent continuously, with a json payload that is an array of log entries, with the log entry format described under Logging for AEM as a Cloud Service. Additional properties are mentioned in the Log Entry Formats section below.
There is also be a property named sourcetype
, which is set to the value aemcdn
.
/.well-known/fastly/logging/challenge
must respond with an asterisk *
in the body and 200 status code.HTTPS AEM logs https-aem
For AEM logs (including apache/dispacher), web requests (POSTs) will be sent continuously, with a json payload that is an array of log entries, with the various log entry formats as described under Logging for AEM as a Cloud Service. Additional properties are mentioned in the Log Entry Formats section below.
There is also be a property named Source-Type
, which is set to one of these values:
- aemaccess
- aemerror
- aemrequest
- aemdispatcher
- aemhttpdaccess
- aemhttpderror
Splunk splunk
kind: "LogForwarding"
version: "1"
metadata:
envTypes: ["dev"]
data:
splunk:
default:
enabled: true
host: "splunk-host.example.com"
token: "${{SPLUNK_TOKEN}}"
index: "aemaacs"
Considerations:
- By default, the port is 443. It can optionally be overridden with a property named
port
. - The sourcetype field will have one of the following values, depending on the specific log: aemaccess, aemerror,
aemrequest, aemdispatcher, aemhttpdaccess, aemhttpderror, aemcdn - If the required IPs have been allowlisted and logs are still not being delivered, verify that there are no firewall rules enforcing Splunk token validation. Fastly performs an initial validation step where an invalid Splunk token is intentionally sent. If your firewall is set to terminate connections with invalid Splunk tokens, the validation process will fail, preventing Fastly from delivering logs to your Splunk instance.
sourcetype
field’s values sent to your Splunk index may have changed, so adjust accordingly.Log Entry Formats log-formats
See Logging for AEM as a Cloud Service for the format of each respective log type (CDN logs, and AEM logs including Apache/Dispatcher).
Since logs from multiple programs and environments may be forwarded to the same logging destination, in addition to the output described in the logging article, the following properties will be included in each log entry:
- aem_env_id
- aem_env_type
- aem_program_id
- aem_tier
As an example, the properties could have the following values:
aem_env_id: 1242
aem_env_type: dev
aem_program_id: 12314
aem_tier: author
Advanced Networking advanced-networking
Some organizations choose to restrict which traffic can be received by the logging destinations.
For the CDN log, you can allow-list the IP addresses, as described in fastly documentation - Public IP List. If that list of shared IP addresses is too large, consider sending traffic to an https server or (non-Adobe) Azure Blob Store where logic can be written to send the logs out of a known IP to their ultimate destination.
For AEM logs (including Apache/Dispatcher), if you have configured advanced networking, you can use the advancedNetworking property to forward them from a dedicated egress IP address or over a VPN.
kind: "LogForwarding"
version: "1"
metadata:
envTypes: ["dev"]
data:
splunk:
default:
enabled: true
host: "splunk-host.example.com"
port: 443
token: "${{SPLUNK_TOKEN}}"
index: "aemaacs"
aem:
advancedNetworking: true
Migrating from Legacy Log Forwarding legacy-migration
Before Log Forwarding configuration was achieved through a self-serve model, customers were requested to open support tickets, where Adobe would initiate the integration.
Customers who had been setup in that manner by Adobe are welcome to adapt to the self-serve model at their convenience. There are several reason to make this transition:
- A new environment (e.g., a new dev env or RDE) has been provisioned.
- Changes to your existing Splunk endpoint or credentials.
- Adobe had setup your log forwarding before CDN logs were available and you would like to receive CDN logs.
- A conscious decision to proactively adapt to the self-serve model so your organization has the knowledge even before a time-sensitive change is necessary.
When ready to migrate, simply configure the YAML file as described in the preceding sections. Use the Cloud Manager config pipeline to deploy to each of the environments where the configuration should be applied.
It is recommended, but not required, that a configuration is deployed to all environments so they are all under self-serve control. If not, you may forget which environments have been configured by Adobe versus those configured in a self-serve way.
sourcetype
field’s values sent to your Splunk index may have changed, so adjust accordingly.