Forward Docker logs (Papertrail)
There are two methods available for shipping container logs from a Docker host to Papertrail. Pick the method that matches how your apps log:
To get a destination and token in Papertrail, add a new log destination. The ingestion token is generated along with the destination. Use the collector endpoints for your account's region:
-
Syslog (TLS):
syslog.collector.<REGION>.cloud.solarwinds.com:6514 -
HTTPS:
https://logs.collector.<REGION>.cloud.solarwinds.com/v1/logs/bulk
Replace <REGION> with the data center your organization uses: na-01, eu-01, or ap-01.
Determining your method
| Option A - logspout | Option B - Fluent Bit | |
|---|---|---|
| Best for | stdout/stderr apps | Apps that also log to files |
| Setup effort | Lowest (one container) | Moderate (config file) |
| Transport | TLS syslog | HTTPS (bulk) |
| File tailing (Y/N) | No | Yes |
| Buffering/retry | Minimal | Yes |
| Per-container System | Yes | Yes (shared host) |
| Per-service Program | Yes (automatic) | Only with one output per service |
Start with Option A for a quick win and correct per-container attribution. Move to Option B when you need file coverage or buffering. For multi-service apps on the HTTP path, use one output per service so Program stays accurate.
How Papertrail fills system and program
The Papertrail Logs Explorer is syslog-shaped and fills two key columns:
-
System (host/source): the
host.nameattribute -
Program (application): the
service.nameattribute
How these columns are populated depends on the transport, which drives the choice between the two options below:
-
Over syslog (Option A), each record carries its own hostname and appname, so System and Program are filled automatically per container.
-
Over HTTP bulk (Option B), host.name and service.name come from an HTTP header that applies to the whole request batch. They are not read from the log body. On a single Docker host, this is fine for System, as every container shares the same host, but Program is constant per output. See the multi-service note below.
Option A: Forward container stdout/stderr with logspout
logspout is a small, stateless container that attaches to the Docker socket. It collects stdout/stderr from your other containers and forwards it over TLS syslog. This is the fastest path for standard containerized apps. Because it sends syslog, each container is attributed correctly—System = host, Program = container—with no extra config.
Docker run
Run the following command in the terminal of your Docker host:
docker run -d \
--name logspout \
--restart=unless-stopped \
--volume=/var/run/docker.sock:/var/run/docker.sock \
-e SYSLOG_STRUCTURED_DATA='solarwinds@41058 token="<SWO_TOKEN>"' \
-e SYSLOG_HOSTNAME="{{.Container.Config.Hostname}}" \
gliderlabs/logspout \
syslog+tls://syslog.collector.<REGION>.cloud.solarwinds.com:6514
Docker Compose
Run the following command in the terminal of your Docker host:
services:
logspout:
image: gliderlabs/logspout
volumes:
- /var/run/docker.sock:/var/run/docker.sock
environment:
SYSLOG_STRUCTURED_DATA: 'solarwinds@41058 token="<SWO_TOKEN>"'
SYSLOG_HOSTNAME: "{{.Container.Config.Hostname}}"
command: syslog+tls://syslog.collector.<REGION>.cloud.solarwinds.com:6514
restart: unless-stopped
Notes
-
Set
SYSLOG_HOSTNAMEso logs show a readable source name instead of the container ID. -
For multi-line events (such as stack traces), use the
multiline+syslog+tls://...prefix on the destination URL. -
logspout collects stdout/stderr only. It does not read log files written inside a container. For file tailing, use Option B: Full coverage with Fluent Bit.
-
See the logspout README for Swarm, environment variables, and filtering options.
Option B: Full coverage with Fluent Bit
Fluent Bit is a collector that can tail both container stdout/stderr and log files inside containers (via Docker volumes or bind mounts), then forward to Papertrail over HTTPS. Use this option when you need file collection, on-disk buffering, retries, or per-source enrichment.
The following examples use a Docker volume (service-logs) for file tailing and a dedicated directory (state) to persist Fluent Bit's internal state across restarts.
Single-host/single-service setup
fluent-bit.conf
Create a file named fluent-bit.conf on your Docker host with the following contents:
[SERVICE]
Flush 2
Log_Level info
Parsers_File /fluent-bit/etc/parsers.conf
# Read Docker stdout/stderr
[INPUT]
Name tail
Path /var/lib/docker/containers/*/*-json.log
Parser docker
Tag myapp
DB /fluent-bit/state/docker.db
# Read application log files
[INPUT]
Name tail
Path /logs/*.log
Tag myapp
DB /fluent-bit/state/files.db
# Send everything to the collector
[OUTPUT]
Name http
Match *
Host logs.collector.<REGION>.cloud.solarwinds.com
Port 443
URI /v1/logs/bulk
Format json_lines
json_date_key timestamp
json_date_format iso8601
Mode tls
tls.verify On
Header Authorization Bearer ${SWO_TOKEN}
Header X-Otel-Resource-Attr host.name=my-host,service.name=my-service
Notes:
-
Use the
/v1/logs/bulkendpoint with thejson_linesformat for newline-batched payloads. -
Combine multiple resource attributes into a single comma-delimited
X-Otel-Resource-Attrheader. -
On a single Docker host, set
host.nameonce in the header. Since every container shares the host, it correctly fills System for every container.
Docker run
Run the following command in the terminal of your Docker host:
docker run -d \ --name fluent-bit \ --restart unless-stopped \ --volume=./fluent-bit.conf:/fluent-bit/etc/fluent-bit.conf:ro \ --volume=./state:/fluent-bit/state \ --volume=/var/lib/docker/containers:/var/lib/docker/containers:ro \ --volume=service-logs:/logs:ro \ fluent/fluent-bit:latest
Docker Compose
Run the following command in the terminal of your Docker host:
services:
myapp:
volumes:
- service-logs:/logs
fluent-bit:
image: fluent/fluent-bit:latest
volumes:
- ./fluent-bit.conf:/fluent-bit/etc/fluent-bit.conf
- ./state:/fluent-bit/state
- /var/lib/docker/containers:/var/lib/docker/containers:ro
- service-logs:/logs:ro
restart: unless-stopped
volumes:
service-logs:
Excluding Fluent Bit's own logs
The example above also forwards Fluent Bit's own logs to Papertrail. If you want to prevent this from happening, you have the following options:
-
Do not produce logs at all. Start the container with
--log-driver=none. -
Produce logs in a different style. Start the container with
--log-driver=syslog. -
Produce and forward only warnings and errors. Use
Log_level warnin Fluent Bit's config. -
Create an exclusion filter in Papertrail.
See the Fluent Bit documentation for installation, inputs, filters, and buffering configuration.
Multi-container and multi-service apps
With a single [OUTPUT], the service.name in the header is applied to every container the collector handles, so the Program column shows one value for the whole host. System is still correct because of the shared host, but Program will not distinguish your services. To attribute Program per service, choose one of the following:
-
Use Option A (logspout): This is the simplest route. Syslog carries each container's name as the appname, so Program is filled per container with no extra config. This is the recommended path for multi-service apps that only need stdout/stderr.
-
One output per service (HTTP): Tag each service's logs at the input, then give each its own
[OUTPUT]with its own header. Because each output's batch contains a single service, the per-batch header is correct.
stdout/stderr
Instead of Docker writing JSON log files that Fluent Bit tails, configure each service to send its logs directly to Fluent Bit. This requires changes to two files.
Modify fluent-bit.conf:
[INPUT]
Name forward
Listen 0.0.0.0
Port 24224
[OUTPUT]
Name http
Match service1
...
Header X-Otel-Resource-Attr host.name=<YOUR_HOST>,service.name=<YOUR_SERVICE_1>
[OUTPUT]
Name http
Match service2
...
Header X-Otel-Resource-Attr host.name=<YOUR_HOST>,service.name=<YOUR_SERVICE_2>
Add a logging section to docker-compose.yml:
services:
service1:
logging:
driver: fluentd
options:
fluentd-address: localhost:24224
tag: service1
fluentd-async: "true"
service2:
logging:
driver: fluentd
options:
fluentd-address: localhost:24224
tag: service2
fluentd-async: "true"
fluent-bit:
image: fluent/fluent-bit:latest
ports:
- "24224:24224"
volumes:
- ./fluent-bit.conf:/fluent-bit/etc/fluent-bit.conf
- ./state:/fluent-bit/state
restart: unless-stopped
Log files
Separate Docker volumes (service1-logs and service2-logs) are used in the example below.
[INPUT]
Name tail
Path /logs/service1/*.log
Tag service1
DB /fluent-bit/state/service1.db
[INPUT]
Name tail
Path /logs/service2/*.log
Tag service2
DB /fluent-bit/state/service2.dbservices:
service1:
volumes:
- service1-logs:/logs
service2:
volumes:
- service2-logs:/logs
fluent-bit:
image: fluent/fluent-bit:latest
volumes:
- ./fluent-bit.conf:/fluent-bit/etc/fluent-bit.conf
- ./state:/fluent-bit/state
- service1-logs:/logs/service1:ro
- service2-logs:/logs/service2:ro
restart: unless-stopped
volumes:
service1-logs:
service2-logs:
All services must be configured to log to the /logs directory.
Multiline parsing
If your application can output JSON (such as using Logback), each log event is emitted as a single JSON object and sent to Papertrail as a single record.
For applications with plain logging, it is necessary to specify a multiline parser depending on the language (multiline.parser) and the presence of Docker parser (the log text itself sits in a field named log):
[FILTER]
Name multiline
Match *
multiline.key_content log
multiline.parser java
For Kubernetes clusters, see Forward Kubernetes logs (Papertrail). A DaemonSet multiplexes many pods, so it must use the syslog output for correct per-pod attribution.
Fallback: Forward via a syslog daemon
If neither logspout nor Fluent Bit can reach the Papertrail collector directly, configure rsyslog on the Docker host to tail the container log files from /var/lib/docker/containers/ and forward them to Papertrail over TLS. The rsyslog template injects the ingestion token as RFC 5424 structured data. See Send logs using syslog (Papertrail) for rsyslog and NXlog configuration.
The scripts are not supported under any SolarWinds support program or service. The scripts are provided AS IS without warranty of any kind. SolarWinds further disclaims all warranties including, without limitation, any implied warranties of merchantability or of fitness for a particular purpose. The risk arising out of the use or performance of the scripts and documentation stays with you. In no event shall SolarWinds or anyone else involved in the creation, production, or delivery of the scripts be liable for any damages whatsoever (including, without limitation, damages for loss of business profits, business interruption, loss of business information, or other pecuniary loss) arising out of the use of or inability to use the scripts or documentation.