HACKING.md

TODO

• account management
  • authenticate via OAuth (Github or Google)
  • RBAC - role based access control
• support different environments (production, jail, staging)
• service management
  • deploy different enviornments
  • view current clusters & environments
  • scale up / scale down
  • automatic and implicit health monitoring
• service health monitoring
  • provide defaults for different kinds of services
  • have a nice Ruby DSL for defining your health check (Luca-O?)
  • have a dedicated daemon running these health checks
    • notifies owners and escalation contacts
• visibility:
  • automatically create graph images to visualize the application architecture
• dashboards
  • custom dashboard available to all members
  • personalised dashboards

Make Me Company independant

• pass endpoints upon docker container startup
• runtime configuration:
  • 1 github organization the users must be part of in order to get in
  • 1+ mesos/marathon cluster endpoint (IP:PORT:ENV)

Names

• serviced web application
• servicectl CLI tool for automation and your favorite shell

Models

User {
  string username;
  string firstname;
  string lastname;
  string email;
  string avatar_url;
  string mobile;
  bool is_admin;
  bool enabled;
}

HealthCheckTemplate {
  string name;            # e.g. "tcp-connect", "smtp", "http", "redis", or ...
  text code;              # ruby code block to execute
  int interval;           # run check every N secs
  int timeout;            # treat check as failed if it runs longer then N secs
}

HealthCheck {
  Service service;
  text code;              # ruby code block to execute
  int interval;           # run check every N secs
  int timeout;            # treat check as failed if it runs longer then N secs
}

Service {
  string name;
  string display_name;
  string docker_image;    # such as example/myapp
  string github_url;      # Github project URL
  Service[] dependencies; # hard dependencies
  Service[] uses;         # soft dependencies (system still up when failing
                          # but degraded)
  Feature[] implements;   # list of features this service [partially] implements
}

# an abstract entity to give a highlevel overview of what is failing due to
# a downtime
Feature {
  string display_name;    # e.g. "Cart Checkout", or "Product Show Page"
  Service[] dependencies; # list of services used to implement this feature
}

Environment {
  string name;            # such as production, jail1, staging1, staging2, ...
  string display_name;
  Instance[] instances;   # list of service instances currently running
  Cluster cluster;        # the reource cluster this env is running on
}

Instance {
  User author;
  Environment env;
  Service service;
  string docker_tag;
  string release_message;
  int instance_count;
  InstanceLocation[] locations;
}

# and if mesos/marathon can provide us the info on which HW node it spawned
# our services, we can list them, too.
# (which we need for individual health checks)
InstanceLocation {
  string hostname;
  int port;
}

Cluster {
  string type;          # such as: "marathon"
  string ip;
  int port;
  Environment[] envs;
}

Configuration {
  string github_org;
  string github_client_id;
  string github_client_secret;
  string marathon_ip;
  int marathon_port;
}