This repository allows you to build a container base image intended to be used to build downstream application images based on a series of modules in order to fast track building of other containers with commonly required features such as:
- Supporting Alpine Linux 3.5-3.22
- Supporting Debian Linux Trixie, Bookworm, Bullseye, Buster
- Multi Processes - Utilize an initialization system to self contain entire application within the image.
- User Management - Update User ID and Group ID permissions dynamically
- Messaging - Send messages from within container
- Monitoring - Track application and container metrics
- Logging - Allow for logrotation, or shipping to centralized log collector
- Scheduling - Execute tasks at specific times
- Secrets Management - Inject secrets or grab from central repository during runtime
- Firewalling - Block or allow remote hosts based on patterns
- VPN - Run VPN clients within the container
- and many customizable settings.
- About
- Maintainer
- Installation
- Configuration
- Developing / Overriding
- Debug Mode
- Shell Access
- Support & Maintenance
- License
Feature limited builds of the image are available on the Github Container Registry To unlock advanced features, one must provide a code to be able to change specific environment variables from defaults. Support the development to gain access to a code.
(container platform) pull ghcr.io/nfrastack/container-base:(image_tag)
Image tag syntax is:
:-<distribution_release><optional_features>
Example:
ghcr.io/nfrastack-container-base:1.0-alpine_3.22_core or
`ghcr.io/nfrastack-container-base:alpine_3.22` or
`ghcr.io/nfrastack-container-base:debian_trixie`
- An otpional tag may exist that matches the CHANGELOG
- Distribution can either be
alpineordebian - Disribution release varies - see the registry for availability
- Features
core- Includesage,cron,logrotate,s6overlay,msmtpfaflopza- Includes everything in core andfail2ban,fluentbit,openbao,zabbix- This is the default image if no features requestedfaflopzazeIncludes everything infaflopzaandzerotier.
Images are built for amd64 by default, with optional support for arm64 and other architectures.
Clone this repository and build the image with your preferred container builder. Customize modules as needed—see the MODULES documentation for details. Building the image from source enables all advanced functionality.
The following directories are used for configuration and can be optionally mapped for persistent storage.
| Directory | Description |
|---|---|
/var/log |
Container, Cron, Zabbix, other log files |
TODO: Add overrides and custom-script info
Below is the complete list of available options that can be used to customize your installation.
- Variables showing an 'x' under the
_FILEcolumn can be used for storing the information inside of a file, useful for secrets. - Variables showing an 'x' under the
Advancedcolumn can only be set if the containers advanced functionality is enabled.
| Parameter | Description | Default | Advanced |
|---|---|---|---|
CONTAINER_NAME |
Used for setting entries in Monitoring and Log Shipping | (hostname) | |
TIMEZONE |
Set Timezone | Etc/GMT |
| Parameter | Description | Default | Advanced |
|---|---|---|---|
CONTAINER_COLORIZE_OUTPUT |
Enable/Disable colorized console output | TRUE |
|
CONTAINER_ENABLE_CUSTOM_BASH_PROMPT |
TODO Set a different bash prompt then '(imagename):(version) HH:MM:SS # ' | x | |
CONTAINER_ENABLE_LOG_TIMESTAMP |
Prefix this images container logs with timestamp | TRUE |
|
CONTAINER_LOG_LEVEL |
Control level of output of container INFO, WARN, NOTICE, DEBUG |
NOTICE |
|
CONTAINER_LOG_PREFIX_TIME_FMT |
Timestamp Time Format | %H:%M:%S |
|
CONTAINER_LOG_PREFIX_DATE_FMT |
Timestamp Date Format | %Y-%m-%d |
|
CONTAINER_LOG_PREFIX_SEPERATOR |
Timestamp seperator | - |
|
CONTAINER_LOG_FILE_LEVEL |
Control level of output of container INFO, WARN, NOTICE, DEBUG |
DEBUG |
x |
CONTAINER_LOG_FILE_NAME |
Internal Container Logs filename | /var/log/container/container.log |
x |
CONTAINER_LOG_FILE_PATH |
Path where to find the internal container logs | /var/log/container/ |
x |
CONTAINER_LOG_FILE_PREFIX_TIME_FMT |
Timestamp Time Format | %H:%M:%S |
x |
CONTAINER_LOG_FILE_PREFIX_DATE_FMT |
Timestamp Date Format | %Y-%m-%d |
x |
CONTAINER_LOG_FILE_PREFIX_SEPERATOR |
Timestamp seperator | - |
x |
This functionality is an attempt to avoid problems within the container that can "run away" from being executed too often.
Sample use cases:
- Alert slack channel when process has executed more than once
- Disable process from executing further if restarted 50 times
- Write to an additional log file..
- Change a file to display "Under Maintenance" on a webserver if this process isn't supposed to be run more than 1 time.
It will pass 5 arguments to a bash script titled the same name as the executing script or if not found, use the default CONTAINER_PROCESS_HELPER_SCRIPT below. Drop your files into the CONTAINER_PROCESS_HELPER_PATH.
For example, if 04-scheduling was starting, it would look for $CONTAINER_PROCESS_HELPER_PATH/04-scheduling and if found execute it while passing the following arguments: DATE,TIME,SCRIPT_NAME,TIMES EXECUTED,HOSTNAME
e.g: 2025-07-01 23:01:04 04-scheduling 2 container
Use the values in your own bash script using the $1 $2 $3 $4 $5 syntax.
Change time and date and settings with these environment variables
| Parameter | Description | Default | Advanced |
|---|---|---|---|
CONTAINER_ENABLE_PROCESS_COUNTER |
Show how many times process has executed in console log | TRUE |
x |
CONTAINER_ENABLE_PROCESS_HELPER |
TRUE |
x | |
CONTAINER_PROCESS_HELPER_PATH |
Path to file external helper scripts | /container/processhelper/ |
x |
CONTAINER_PROCESS_HELPER_SCRIPT |
Default helper script name | processhelper.sh |
x |
CONTAINER_PROCESS_HELPER_DATE_FMT |
Date format passed to external script | %Y-%m-%d |
x |
CONTAINER_PROCESS_HELPER_TIME_FMT |
Time format passed to external script | %H:%M:%S |
x |
CONTAINER_PROCESS_RUNAWAY_PROTECTOR |
Disables a service if executed more than (x) amount of times | TRUE |
x |
CONTAINER_PROCESS_RUNAWAY_DELAY |
Delay in seconds to restart process | 1 |
x |
CONTAINER_PROCESS_RUNAWAY_LIMIT |
The amount of times it needs to restart before disabling | 10 |
x |
CONTAINER_PROCESS_RUNAWAY_SHOW_OUTPUT_FINAL |
Show the program Output on the final execution before disabling | TRUE |
x |
Only valid if appropriate module has been installed within the image. See MODULES documentation for more information to see other configurable options.
| Parameter | Description | Default | Advanced |
|---|---|---|---|
CONTAINER_ENABLE_METRICS |
Enable Metrics | TRUE |
x |
CONTAINER_ENABLE_MONITORING |
Enable Monitoring of applications or metrics | TRUE |
|
CONTAINR_MONITORING_BACKEND |
What monitoring agent to use zabbix |
zabbix |
|
CONTAINER_ENABLE_MESSAGING |
Enable Messaging services like SMTP | TRUE |
|
CONTAINER_MESSAGING_BACKEND |
Messaging Backend - presently only msmtp |
msmtp |
|
CONTAINER_ENABLE_SCHEDULING |
Enable Scheduled Tasks | TRUE |
|
CONTAINR_SCHEDULING_BACKEND |
What scheduling tool to use cron |
cron |
|
CONTAINER_ENABLE_LOGROTATE |
Enable Logrotate (if scheduling enabled) | TRUE |
|
CONTAINER_ENABLE_LOGSHIPPING |
Enable Log Shipping | FALSE |
|
CONTAINER_LOGSHIPPING_BACKEND |
Log shipping backend fluent-bit |
fluent-bit |
|
CONTAINER_ENABLE_FIREWALL |
Enable Firewall Functionality | FALSE |
|
CONTAINER_FIREWALL_BACKEND |
What Firewall backend to use iptables |
iptables |
Included when proper capabilities are set on image is the capability to set up detailed block / allow rules via a firewall on container start.
Presently only iptables is supported.
You must use run your containers with the following capabilities added: NET_ADMIN, NET_RAW
| Parameter | Description | Default |
|---|---|---|
CONTAINER_ENABLE_FIREWALL |
Enable Firewall Functionality | FALSE |
CONTAINER_FIREWALL_BACKEND |
What Firewall backend to use iptables |
iptables |
FIREWALL_RULE_00 |
Firewall rule to execute | |
FIREWALL_RULE_01 |
Next firewall rule to execute |
One can use the FIREWALL_RULE_XX environment variables to pass rules to the firewall. In this example I am going to block someone from being able to access a port except if from a specific IP address:
FIREWALL_RULE_00=-I INPUT -p tcp -m tcp -s 101.69.69.101 --dport 389 -j ACCEPT
FIREWALL_RULE_01=-I INPUT -p tcp -m tcp -s 0.0.0.0/0 --dport 389 -j DROPInstead of relying on environment variables one can put a iptables-restore compatible ruleset below and it will be imported on container start.
| Parameter | Description | Default |
|---|---|---|
IPTABLES_RULES_PATH |
Path for IPTables Rules | /contianer/iptables/ |
IPTABLES_RULES_FILE |
IPTables Rules File to restore if exists on container start | iptables.rules |
Sometimes you may need to add aliases or change the response of DNS lookups.. This will add an entry to the containers hosts file.
| Parameter | Description |
|---|---|
CONTAINER_HOST_OVERRIDE_01 |
Create manual hosts entry |
Make the value <destination> override1 override2 eg 1.2.3.4 example.org example.com.
If you omit an IP Address and instead use a domain name it will attempt to look it up to an IP eg proxy example.com example.org
This image can execute commands or scripts after the container initializes so that you can perform commands without having to rebuild the entire image or maintain a new codebase. Simply pass the command you wish to execute, or execute a script from the location from a mapped volume.
| Parameter | Description | Default | Advanced |
|---|---|---|---|
CONTAINER_CUSTOM_PATH |
Used for adding custom files into the image upon startup | /override/custom |
x |
CONTAINER_CUSTOM_SCRIPTS_PATH |
Used for adding custom scripts to execute upon startup | /override/custom-scripts |
x |
CONTAINER_POST_INIT_COMMAND |
If you wish to execute a command in the container after all services have initialized | ||
| enter it here. Seperate multiple by commas. | |||
CONTAINER_POST_INIT_SCRIPT |
If you wish to execute a script in the container after all services have initialized | ||
| enter it here. Seperate multiple by commas. |
If you wish to change the internal id for users and groups you can set environment variables to do so.
e.g. If you add USER_NGINX=1000 it will reset the containers nginx user id from 82 to 1000 -
If you enable DEBUG_PERMISSIONS=TRUE all the users and groups have been modified in accordance with
environment variables will be displayed in output.
Hint, also change the Group ID to your local development users UID & GID and avoid Docker permission issues when developing.
| Parameter | Description |
|---|---|
CONTAINER_USER_<USERNAME> |
The user's UID in /etc/passwd will be modified with new UID |
CONTAINER_GROUP_<GROUPNAME> |
The group's GID in /etc/group and /etc/passwd will be modified with new GID |
CONTAINER_GROUP_ADD_<GROUPNAME> |
The username will be added in /etc/group after the group name defined |
This base image has been used over a hundred times to successfully build secondary images. My methodology is admittedly straying from the "one process per container" rule, however this methodology allows me to put together images at a rapid pace, and if more complex scalability is required, the work is split into their own individual images. Since you are reading this here's a crash course at how the image works - (WIP):
See /container/base/functions/ for more detailed documentation for the various commands and functions/shortcuts
- Put defaults in
/container/defaults/(script name) - Put functions in
/container/functions/(script name) - Put Initialization script in
container/init/init.d/(script name) - Put at the top:
#!/command/with-contenv bash # Pull in Container Environment Variables from Containerfile/Container Runtime
source /container/base/functions/container/init # Pull in all custom container functions from this image
prepare_service single # Read functions and defaults only from files matching this script filename - see detailed docs for more
SERVICE_NAME="nginx" # set the prefix for any logging
.. your scripting ..
print_info "This an INFO log"
print_warn "This a WARN log"
print_error "This is a ERROR log"
liftoff # this writes to the state files at /container/state/.container/ to prove the script executed properlyOnce initialization has completed, it will then execute a service script.
-
Put the services script in
/container/run/available/(script_name)/runPut at the top:
#!/command/with-contenv bash # Pull in Container Environment Variables from Containerfile/Container Runtime
source /container/base/functions/container/init # Pull in all custom container functions from this image
SERVICE_NAME="nginx" # set the prefix for any logging
check_container_initialized # Check to make sure that the container properly initialized before proceeding
check_service_initialized init # Check to see if the init/(script name) executed correctly, otherwise wait until it has done
liftoff # Prove script was able to execute properly
print_start "Starting processname" # Show STARTING log prefix, and also show if enabled a counter, and execute process watchdog script
fakeprocess (args) # whatever your process you want to start is| Parameter | Description | Default |
|---|---|---|
CONTAINER_SKIP_SANITY_CHECK |
Skip the checking to see if all scripts in /container/init/init.d executed correctly | FALSE |
DEBUG_MODE |
Show all script output (set -x) | FALSE |
SERVICE_NAME |
Used for prefixing the script that is running | container |
When using this as a base image, create statements in your startup scripts to check for existence of DEBUG_MODE=TRUE
and set various parameters in your applications to output more detail, enable debugging modes, and so on.
In this base image it does the following:
- Shows all script output (equivalent to set -x)
For debugging and maintenance, bash and sh are available in the container.
- For community help, tips, and community discussions, visit the Discussions board.
- For personalized support or a support agreement, see Nfrastack Support.
- To report bugs, submit a Bug Report. Usage questions may be closed as not-a-bug.
- Feature requests are welcome, but not guaranteed. For prioritized development, consider a support agreement.
- Updates are best-effort, with priority given to active production use and support agreements.
This project is licensed under the MIT License - see the LICENSE file for details.