TODO
- Docker
- Docker compose
- CI
Notification system is a system serve for notification a Clients with some messages they want to receive from somewhere
Functional requirements answer the questions:
-
"What does the system do?"
-
"Why does the system do this?"
-
Notify a users according configuration they have been set
-
A frontend with a configuration for notification configuration
The flow is a Client set message, time and channel to receive a notification, all this stuff will validate and registered in Database, and then send to Client in a proper time.
For notification I am going to use push model. Notifications methods are two - email and telegram messages.
Maybe other like whatsapp or SMS.
Non-functional requirements answer the questions:
- "How well does the system work?"
- "Under what conditions should it function?"
Also there is necessary to clear database after if notification is not set to repeated.
Need to control deduplication, with timeouts.
When a particular message has been sent, need to register it like send, maybe in other DB, but check status first. If some messages cant be sent for some reason need to resend it or notify developers or system administrators to check the situatuon
All the notification has a priority, not all of them a important. So, one of them has to be send faster and other later. Some notification can be send in email, some with a messangers like tg or whatsapp.
For example:
| Priority | Example | Default |
|---|---|---|
critical |
Security alerts | Messengers |
high |
Mentions, direct messages | Messengers |
medium |
Comments, reactions | |
low |
Marketing, digest |
There are some components of a system:
- Data
- Processing
- RabbitMQ
- Writer
- Scheduler
- Redis
- Relay
- Notifier
It is a data sources from where you want to get data for a notification and also the time configurations how and when you want to receive it
For now it will some kind of table but maybe it is good idea configure it with a frontend web app
Output: file (or JSON)
A Web Application which get data from user
Parse and validate JSON file then put it in queue named QID (queue input data).
Use Flask like a web server
Output: JSON
In fact it is a consumer which get data from QID and put it in PG
This app check last actual records in DB and put them into Redis then delete them to free a database
Contain one or more stores to save some data from DB
Get data from Redis and put them in queues QOT or QOM
Consumer which get messages from RabbitMQ, from QOT (queue out telegram) or QOM (queue out mail) and send it for a correspond client
Message Queue Design
| Abbreviation | Full Name | Purpose |
|---|---|---|
| QID | Queue Input Data | Queue for incoming data |
| QOT | Queue Out Telegram | Queue for outgoing Telegram messages |
| QOM | Queue Out Mail | Queue for outgoing email messages |
Contains two tables for clients to notify:
- Notifications, this one contain simple record for one-time notification
| Field | Description |
|---|---|
id |
Entity ID |
client_name |
Real name or nickname |
client_tg_nick |
Telegram nickname of a client |
n_method |
Notification method: Telegram or Email |
n_text |
Text of the notification message |
n_time |
Time when the notification should be sent |
r_time |
Time when the notification has been received |
prioriy |
Priority of a notifications |
is_repeat |
Whether the notification should be repeated |
- Notifications_Repeated, this one contain simple record for repeated notification
| Field | Description |
|---|---|
id |
Entity ID |
client_name |
Real name or nickname |
client_tg_nick |
Telegram nickname of a client |
n_method |
Notification method: Telegram or Email |
n_text |
Text of the notification message |
n_time |
Time when the notification should be sent |
r_time |
Time when the notification has been received |
prioriy |
Priority of a notifications |
flowchart TB
subgraph Processing["Processing"]
PWA[WebApp]
PVA[Validator]
end
subgraph RabbitMQ["RabbitMQ"]
QID@{ shape: das, label: "Queue_QID" }
QOT@{ shape: das, label: "Queue_QOT" }
QOM@{ shape: das, label: "Queue_QOM" }
end
subgraph Writer["Writer"]
WRR[Writer]
end
subgraph Relay["Relay"]
RLY[Relay]
end
subgraph Scheduler["Scheduler"]
SCR[Scheduler]
end
subgraph PostgreSQL["PostgreSQL"]
direction TB
DBN[(Notifications)]
DBR[(Notifications_Repeated)]
DBN ~~~ DBR
end
subgraph Notifier["Notifier"]
Email[📧 Email]
Telegram[💬 Telegram]
end
Sources[(Sources)] --> PWA --> PVA --> QID --> WRR --> PostgreSQL
SCR --> PostgreSQL
SCR --> Redis[(Redis)]
RLY --> Redis[(Redis)]
RLY --> QOM --> Email
RLY --> QOT --> Telegram