Most of these tools use platformsh-cli. You can install platform-cli inside your container like this:
https://docs.platform.sh/development/cli/api-tokens.html#install-the-cli-on-a-platformsh-environment
hooks:
build: |
curl -sS https://platform.sh/cli/installer | php
Download the tools and commit them to your repository. You can modify it as you like. https://raw.githubusercontent.com/matthiaz/platformsh-tools/master/block_ddos.sh
Some of these tools can be triggered using cron scripts as described here https://docs.platform.sh/configuration/app/cron.html#cron-jobs Make sure you set the correct run time. https://crontab.guru/ can be of assistance.
This script can be used as a somewhat replacement to the pecl install <package>
package manager that is used by php to compile and install custom extensions.
It doesn't do search, but can compile packages just fine.
Usage:
Download the sourcecode: https://raw.githubusercontent.com/matthiaz/platformsh-tools/master/pecl and place it in your project root.
Then add this as a hook:
hooks:
build: |
set -e
chmod +x pecl
./pecl install grpc
It will automatically add a /app/php.ini with the extension requested. First builds will be slow since it compiles the package, but subsequent builds will be cached in PLATFORM_CACHE_DIR automatically.
This script can be used to automatically block ip addresses using the platform environment:http-access
tool.
*IMPORTANT NOTE: This will automatically redeploy your environment. *
#./block_ddos.sh [MAX_ALLOWED_REQUESTS=60] [PERIOD='last minute']
#./block_ddos.sh
#./block_ddos.sh 60
#./block_ddos.sh 60 'last minute'
#./block_ddos.sh 3600 'last hour'
#./block_ddos.sh 3600 'now -1hour'
-
bash block_ddos.sh
- Looks at /var/log/access.log
- Blocks all IP's that were seen more than 60 times in the previous minute.
-
bash block_ddos.sh 300
- Looks at /var/log/access.log
- Blocks all IP's that were seen more than 300 times in the previous minute.
You could run the script manually or you could put it in a cron that runs every 15 minutes or so.
e.g.:
hooks:
build: |
curl -sS https://platform.sh/cli/installer | php
crons:
blockddos:
spec: '*/15 * * * *' #every x minutes
cmd: |
if [ "$PLATFORM_BRANCH" = master ]; then
bash block_ddos.sh 1000 'now -15minutes'
fi
This script can be used to analyse your access.log and give a good estimation on how much data is being sent. This can be used to determine the right size of a CDN solution (CloudFront, Fastly,...). It is a simple bash script and can be used without needing a platform-cli API key.
NOTE: We trim log files, so only the last 100MB of log files will be parsed (usually a few days worth)
#./bandwith_stats.sh
curl -L -sS https://raw.githubusercontent.com/matthiaz/platformsh-tools/master/bandwith_stats.sh | bash
Or if you want to download it
curl -L -sS --output /tmp/bandwith_stats.sh https://raw.githubusercontent.com/matthiaz/platformsh-tools/master/bandwith_stats.sh && bash /tmp/bandwith_stats.sh
You don't have to download it, you can simply pipe the script straight to curl as per the first example.
This script can be used to qucikly add lots of env variables without having to wait for each deploy
git clone the repository, you can then run
bash add_multiple_variables.sh PROJECT_ID ENVIRONMENT_TO_REDEPLOY 'var1=val1,var2=val2,var3=val3' VARIABLE_LEVEL (project or environment=default)
bash add_multiple_variables.sh xj2nccsddc57w master 'var1=val1,var2=val2,var3=val3' project
project level variables
bash add_multiple_variables.sh xj2nccsddc57w master 'var1=val1,var2=val2,var3=val3' project
environment level variables
This script can be used to qucikly add lots of domains without having to wait for each deploy
git clone the repository, you can then run
`./add_multiple_domains.sh PROJECT_ID ENVIRONMENT_TO_REDEPLOY 'domain1.com,domain2.com,domain3.com'
./add_multiple_domains.sh tug2vhb33pje6 master 'test001.giveatree.world,test002.giveatree.world,test003.giveatree.world,test004.giveatree.world,test005.giveatree.world,test006.giveatree.world'
This script can be used to quickly delete lots of domains without having to wait for each deploy
git clone the repository, you can then run
`./delete_multiple_domains.sh PROJECT_ID ENVIRONMENT_TO_REDEPLOY 'domain1.com,domain2.com,domain3.com'
./delete_multiple_domains.sh tug2vhb33pje6 master 'test001.giveatree.world,test002.giveatree.world,test003.giveatree.world,test004.giveatree.world,test005.giveatree.world,test006.giveatree.world'
This script can be used to quickly delete lots of users without having to wait for each deploy
git clone the repository, you can then run
`./delete_multiple_users.sh PROJECT_ID ENVIRONMENT_TO_REDEPLOY 'user1@example.com,user2@example.com,user3@example.com'
./delete_multiple_users.sh tug2vhb33pje6 master 'test001@giveatree.world,test002@giveatree.world,test003@giveatree.world'
brew.sh is a package manager specifically for macos, but it can also be used to install packages on linux. This allows you to install all sorts of things fairly easy within in the build hook.
All you have to do is copy the install_brew_packages.sh file into the root of your project (where your .platform.app.yaml
lives) and then you can call it in the build hook. The script takes 1 or more arguments. Simply specify the packages you want to install.
Building from source with brew is rather slow, but the script will automatically use the cache folder. This will ensure that the subsequent builds are fast.
To install duf
and lnav
you can do this:
hooks:
build: |
set -e
bash install_brew_packages.sh duf lnav
This script allows you to syncronize a database table with banned ips, with your fastly CDN it allows you to easily ban_ips from your application. Prerequisites:
- environment variables FASTLY_API_TOKEN and FASTLY_SERVICE_ID need to be set (should be done already but double check)
- ensure that you change TABLE_TO_GET_IPS_FROM in the script to the correct table name
Triggering it can be done manually
bash ./autoban_in_fastly
Or via a cron script every x minutes
crons:
autoban:
spec: '*/13 * * * *'
commands:
start: ./autoban_in_fastly.sh
shutdown_timeout: 31
The add_ai_bots_to_robots_txt.sh script appends known AI bots to the robots.txt and can simply be added to your build hook.
hooks:
build: |
set -e
bash add_ai_bots_to_robots_txt.sh
AI is cool and all, but many of these AI bots use very agressive crawlers that actively take down websites due to the amount of requests they send. Even if your site is able to handle the requests, you are paying server resources to feed them content that you created and they give nothing back. They are not search engine bots, they will not be sending traffic to your site in any way in the future.
Many customers have asked to block them. And you can, but one quick easy win is to ensure that your robots.txt file is up to date and tells them correctly not to browse your site.
get_resources.sh helps with diagnosing how much resources are allocated to all containers.
Usage: get_resources.sh <projectid> <environment>
Example:
bash get_resources.sh y3dnj3qm242nq master
Service cpu mem disk
------- --- --- ----
drupal 0.40 128 12021
cache 0.08 256 0
db 0.25 288 15045
------- --- --- ----
Total 0.73 672 27066
Plan:
production: { legacy_development: false, max_cpu: 0.96, max_memory: 768, max_environments: 1 }