// This is the pm2 configuration file for npms-api
{
    apps: [
        {
            name: 'npms-api',
            script: './server.js',
            args: [],
            instances: 3,
            env : { NODE_ENV: 'production' },
            max_memory_restart: '2018M',  // Restart if it's getting close to the limit
            node_args: '--max_old_space_size=2048',
            pmx: false,  // We were experiecing PMX has a lot of memory leaks, see: https://github.com/keymetrics/pmx/issues/23
        },
    ],
    deploy: {
        production: {
            user: 'www',
            host: '163.172.160.79',
            ref: 'origin/master',
            repo: 'https://github.com/npms-io/npms-api.git',
            path: '/home/www/npms-api',
            'post-deploy': '\
cp ../local.json5 ./config/ && \
npm update --loglevel http --production && \
pm2 startOrRestart ecosystem.json5 --env production',
        },
    },
}


Enable pmx again, leak was caused by an external dep that is not used here.
// This is the pm2 configuration file for npms-api
{
    apps: [
        {
            name: 'npms-api',
            script: './server.js',
            args: [],
            instances: 3,
            env : { NODE_ENV: 'production' },
            max_memory_restart: '2018M',  // Restart if it's getting close to the limit
            node_args: '--max_old_space_size=2048',
        },
    ],
    deploy: {
        production: {
            user: 'www',
            host: '163.172.160.79',
            ref: 'origin/master',
            repo: 'https://github.com/npms-io/npms-api.git',
            path: '/home/www/npms-api',
            'post-deploy': '\
cp ../local.json5 ./config/ && \
npm update --loglevel http --production && \
pm2 startOrRestart ecosystem.json5 --env production',
        },
    },
}