1
0
mirror of https://github.com/taigrr/nats.docs synced 2025-01-18 04:03:23 -08:00

Reworded to describe behaviour of non java clients and merge examples.

Specifically removed the reference to flush which is not influenced by
this setting.
Merged the examples as both values together make more sense.

Signed-off-by: Matthias Hanel <mh@synadia.com>
This commit is contained in:
Matthias Hanel 2020-02-03 15:39:04 -05:00
parent 92e62d1282
commit bf9b8a8247

View File

@ -1,18 +1,16 @@
# Ping/Pong Protocol
The client and server use a simple PING/PONG protocol to check that they are both still connected. The client will ping the server on a regular, configured interval so that the server usually doesn't have to initiate the PING/PONG interaction.
The client and server use a simple PING/PONG protocol to check that either of them are still connected to the other. On a regular interval he client will ping the server, which responds with a pong. Once a configurable maximum of outstanding pings without a single pong reply is hit, the connection is closed as stale. Together these two values define a timeout for the connection. In the pressence of traffic, such as messages or client side pings, the server will not initiate the PING/PONG interaction.
![](../../.gitbook/assets/pingpong.svg)
## Set the Ping Interval
If you have a connection that is going to be open a long time with few messages traveling on it, setting this PING interval can control how quickly the client will be notified of a problem. However on connections with a lot of traffic, the client will often figure out there is a problem between PINGS, and as a result the default PING interval is often on the order of minutes. To set the interval to 20s:
If you have a connection that is going to be open a long time with few messages traveling on it, setting the PING interval and/or limit how many are outstanding, can control how quickly the client will be notified of a problem. However on connections with a lot of traffic, the client will often figure out there is a problem between PINGS, and as a result the default PING interval is often on the order of minutes. To set the interval to 20s and limit outstanding pings to 5, thus force a closed connection after 100s of inactivity:
{% tabs %}
{% tab title="Go" %}
```go
// Set Ping Interval to 20 seconds
nc, err := nats.Connect("demo.nats.io", nats.Name("API Ping Example"), nats.PingInterval(20*time.Second))
nc, err := nats.Connect("demo.nats.io", nats.Name("API Ping Example"), nats.PingInterval(20*time.Second), nats.MaxPingsOutstanding(5))
if err != nil {
log.Fatal(err)
}
@ -27,93 +25,6 @@ defer nc.Close()
Options options = new Options.Builder().
server("nats://demo.nats.io:4222").
pingInterval(Duration.ofSeconds(20)). // Set Ping Interval
build();
Connection nc = Nats.connect(options);
// Do something with the connection
nc.close();
```
{% endtab %}
{% tab title="JavaScript" %}
```javascript
let nc = NATS.connect({
pingInterval: 20*1000, //20s
url: "nats://demo.nats.io:4222"
});
```
{% endtab %}
{% tab title="Python" %}
```python
nc = NATS()
await nc.connect(
servers=["nats://demo.nats.io:4222"],
# Set Ping Interval to 20 seconds
ping_interval=20,
)
# Do something with the connection.
```
{% endtab %}
{% tab title="Ruby" %}
```ruby
require 'nats/client'
NATS.start(ping_interval: 20) do |nc|
nc.on_reconnect do
puts "Got reconnected to #{nc.connected_server}"
end
nc.on_disconnect do |reason|
puts "Got disconnected! #{reason}"
end
# Do something with the connection
end
```
{% endtab %}
{% tab title="TypeScript" %}
```typescript
// will throw an exception if connection fails
let nc = await connect({
pingInterval: 20*1000, //20s
url: "nats://demo.nats.io:4222"
});
nc.close();
```
{% endtab %}
{% endtabs %}
## Limit Outgoing Pings
The PING/PONG interaction is also used by most of the clients as a way to flush the connection to the server. Clients that cache outgoing messages provide a flush call that will run a PING/PONG. The flush will wait for the PONG to return, telling it that all cached messages have been processed, including the PING. The number of cached PING requests can be limited in most clients to insure that traffic problems are identified early. This configuration for _max outgoing pings_ or similar will usually default to a small number and should only be increased if you are worried about fast flush traffic, perhaps in multiple threads.
For example, to set the maximum number of outgoing pings to 5:
{% tabs %}
{% tab title="Go" %}
```go
// Set maximum number of PINGs out without getting a PONG back
// before the connection will be disconnected as a stale connection.
nc, err := nats.Connect("demo.nats.io", nats.Name("API MaxPing Example"), nats.MaxPingsOutstanding(5))
if err != nil {
log.Fatal(err)
}
defer nc.Close()
// Do something with the connection
```
{% endtab %}
{% tab title="Java" %}
```java
Options options = new Options.Builder().
server("nats://demo.nats.io:4222").
maxPingsOut(5). // Set max pings in flight
build();
Connection nc = Nats.connect(options);
@ -127,6 +38,7 @@ nc.close();
{% tab title="JavaScript" %}
```javascript
let nc = NATS.connect({
pingInterval: 20*1000, //20s
maxPingOut: 5,
url: "nats://demo.nats.io:4222"
});
@ -139,10 +51,9 @@ nc = NATS()
await nc.connect(
servers=["nats://demo.nats.io:4222"],
# Set maximum number of PINGs out without getting a PONG back
# before the connection will be disconnected as a stale connection.
# Set Ping Interval to 20 seconds
ping_interval=20,
max_outstanding_pings=5,
ping_interval=1,
)
# Do something with the connection.
@ -153,7 +64,7 @@ await nc.connect(
```ruby
require 'nats/client'
NATS.start(max_outstanding_pings: 5) do |nc|
NATS.start(ping_interval: 20, max_outstanding_pings: 5) do |nc|
nc.on_reconnect do
puts "Got reconnected to #{nc.connected_server}"
end
@ -171,6 +82,7 @@ end
```typescript
// will throw an exception if connection fails
let nc = await connect({
pingInterval: 20*1000, //20s
maxPingOut: 5,
url: "nats://demo.nats.io:4222"
});
@ -178,4 +90,3 @@ nc.close();
```
{% endtab %}
{% endtabs %}