Skip to content
This repository has been archived by the owner on Aug 23, 2023. It is now read-only.

Commit

Permalink
Merge pull request #394 from raintank/removeKafkaMdam
Browse files Browse the repository at this point in the history
remove kafkamdam input plugin
  • Loading branch information
Dieterbe authored Nov 17, 2016
2 parents a036c09 + cec9103 commit 82af80b
Show file tree
Hide file tree
Showing 7 changed files with 2 additions and 220 deletions.
13 changes: 0 additions & 13 deletions docs/config.md
Original file line number Diff line number Diff line change
Expand Up @@ -203,19 +203,6 @@ consumer-max-processing-time = 1s
net-max-open-requests = 100
```

### kafka-mdam input (optional, discouraged)

```
[kafka-mdam-in]
enabled = false
# tcp address (may be given multiple times as a comma-separated list)
brokers = kafka:9092
# kafka topic (may be given multiple times as a comma-separated list)
topics = mdam
# consumer group name
group = group1
```

## clustering transports ##
### kafka as transport for clustering messages (recommended)

Expand Down
9 changes: 1 addition & 8 deletions docs/inputs.md
Original file line number Diff line number Diff line change
Expand Up @@ -24,14 +24,7 @@ note: it does not implement [carbon2.0](http://metrics20.org/implementations/)
## Kafka-mdm (recommended)

`mdm = MetricData Messagepack-encoded` [MetricData schema definition](https://github.com/raintank/schema/blob/master/metric.go#L20)
This is a kafka input wherein each point is sent as a unique kafka message. This is the best way,
even though we haven't gotten it to perform on par with kafka-mdam yet, but we expect to get there soon.
This is a kafka input wherein each point is sent as a unique kafka message.
This is the recommended input option if you want a queue.

## Kafka-mdam (experimental, discouraged)

`mdm = MetricDataArray Messagepack-encoded` [MetricDatayArray schema definition](https://github.com/raintank/schema/blob/master/metric.go#L47)
This is a kafka input that uses application-level batches stored within single kafka messages.
It is discouraged because this does not allow proper routing/partitioning of messages and will be removed.
It only exists to compare performance numbers against kafka-mdm, and make mdm as fast as mdam.
It also doesn't have the performance tuning options and offset control that has kafka-mdm has. (kafka-mdam always uses latest offset).
157 changes: 0 additions & 157 deletions input/kafkamdam/kafkamdam.go

This file was deleted.

11 changes: 0 additions & 11 deletions metrictank-sample.ini
Original file line number Diff line number Diff line change
Expand Up @@ -172,17 +172,6 @@ consumer-max-processing-time = 1s
# How many outstanding requests a connection is allowed to have before sending on it blocks
net-max-open-requests = 100

### kafka-mdam input (optional, discouraged)
[kafka-mdam-in]
enabled = false
# tcp address (may be given multiple times as a comma-separated list)
brokers = kafka:9092
# kafka topic (may be given multiple times as a comma-separated list)
topics = mdam
# consumer group name
group = group1


## clustering transports ##

### kafka as transport for clustering messages (recommended)
Expand Down
10 changes: 1 addition & 9 deletions metrictank.go
Original file line number Diff line number Diff line change
Expand Up @@ -29,7 +29,6 @@ import (
"github.com/raintank/metrictank/idx/memory"
"github.com/raintank/metrictank/input"
inCarbon "github.com/raintank/metrictank/input/carbon"
inKafkaMdam "github.com/raintank/metrictank/input/kafkamdam"
inKafkaMdm "github.com/raintank/metrictank/input/kafkamdm"
"github.com/raintank/metrictank/mdata"
"github.com/raintank/metrictank/mdata/chunk"
Expand Down Expand Up @@ -162,7 +161,6 @@ func main() {
// load config for metric ingestors
inCarbon.ConfigSetup()
inKafkaMdm.ConfigSetup()
inKafkaMdam.ConfigSetup()

// load config for cluster handlers
notifierNsq.ConfigSetup()
Expand Down Expand Up @@ -214,12 +212,11 @@ func main() {

inCarbon.ConfigProcess()
inKafkaMdm.ConfigProcess(*instance)
inKafkaMdam.ConfigProcess(*instance)
notifierNsq.ConfigProcess()
notifierKafka.ConfigProcess(*instance)
api.ConfigProcess()

if !inCarbon.Enabled && !inKafkaMdm.Enabled && !inKafkaMdam.Enabled {
if !inCarbon.Enabled && !inKafkaMdm.Enabled {
log.Fatal(4, "you should enable at least 1 input plugin")
}

Expand Down Expand Up @@ -315,11 +312,6 @@ func main() {
inputs = append(inputs, inKafkaMdm.New(stats))
}

if inKafkaMdam.Enabled {
sarama.Logger = l.New(os.Stdout, "[Sarama] ", l.LstdFlags)
inputs = append(inputs, inKafkaMdam.New(stats))
}

accountingPeriod := dur.MustParseUNsec("accounting-period", *accountingPeriodStr)

metrics = mdata.NewAggMetrics(store, chunkSpan, numChunks, chunkMaxStale, metricMaxStale, ttl, gcInterval, finalSettings)
Expand Down
11 changes: 0 additions & 11 deletions scripts/config/metrictank-docker.ini
Original file line number Diff line number Diff line change
Expand Up @@ -167,17 +167,6 @@ consumer-max-processing-time = 1s
# How many outstanding requests a connection is allowed to have before sending on it blocks
net-max-open-requests = 100

### kafka-mdam input (optional, discouraged)
[kafka-mdam-in]
enabled = false
# tcp address (may be given multiple times as a comma-separated list)
brokers = kafka:9092
# kafka topic (may be given multiple times as a comma-separated list)
topics = mdam
# consumer group name
group = group1


## clustering transports ##

### kafka as transport for clustering messages (recommended)
Expand Down
11 changes: 0 additions & 11 deletions scripts/config/metrictank-package.ini
Original file line number Diff line number Diff line change
Expand Up @@ -155,17 +155,6 @@ consumer-max-processing-time = 1s
# How many outstanding requests a connection is allowed to have before sending on it blocks
net-max-open-requests = 100

### kafka-mdam input (optional, discouraged)
[kafka-mdam-in]
enabled = false
# tcp address (may be given multiple times as a comma-separated list)
brokers = localhost:9092
# kafka topic (may be given multiple times as a comma-separated list)
topics = mdam
# consumer group name
group = group1


## clustering transports ##

### kafka as transport for clustering messages (recommended)
Expand Down

0 comments on commit 82af80b

Please sign in to comment.