Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Ipmi.v2.schema #4450

Merged
merged 4 commits into from
Jul 31, 2018
Merged

Ipmi.v2.schema #4450

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
143f6c3
Adding V2 schema for IPMI_sensor
jhg03a Jan 23, 2018
09e834f
Go-metalinter cleanup
jhg03a Jan 28, 2018
7f1e60c
Refactored to use named arrays instead of indexes for improved readab…
jhg03a Jan 28, 2018
c2a5a54
Addressing feedback on PR #4450
jhg03a Jul 31, 2018
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions etc/telegraf.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1976,6 +1976,9 @@
# ## Timeout for the ipmitool command to complete
# timeout = "20s"

# ## Schema Version: (Optional, defaults to version 1)
# schemaVersion = 2


# # Gather packets and bytes counters from Linux ipsets
# [[inputs.ipset]]
Expand Down
63 changes: 48 additions & 15 deletions plugins/inputs/ipmi_sensor/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,10 @@ If no servers are specified, the plugin will query the local machine sensor stat
```
ipmitool sdr
```
or with the version 2 schema:
```
ipmitool sdr elist
```

When one or more servers are specified, the plugin will use the following command to collect remote host sensor stats:

Expand Down Expand Up @@ -41,19 +45,36 @@ ipmitool -I lan -H SERVER -U USERID -P PASSW0RD sdr

## Timeout for the ipmitool command to complete. Default is 20 seconds.
timeout = "20s"

## Schema Version: (Optional, defaults to version 1)
schemaVersion = 2
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you change this to be metric_version, this will match the option I used in in the mysql plugin for similar functionality. Make sure to use snake_case too.

```

### Measurements

Version 1 schema:
- ipmi_sensor:
- tags:
- name
- unit
- host
- server (only when retrieving stats from remote servers)
- fields:
- status (int)
- status (int, 1=ok status_code/0=anything else)
- value (float)

Version 2 schema:
- ipmi_sensor:
- tags:
- name
- entity_id (can help uniquify duplicate names)
- status_code (two letter code from IPMI documentation)
- status_desc (extended status description field)
- unit (only on analog values)
- host
- server (only when retrieving stats from remote)
- fields:
- value (float)

#### Permissions

Expand All @@ -68,24 +89,36 @@ KERNEL=="ipmi*", MODE="660", GROUP="telegraf"

### Example Output

#### Version 1 Schema
When retrieving stats from a remote server:
```
ipmi_sensor,server=10.20.2.203,unit=degrees_c,name=ambient_temp status=1i,value=20 1458488465012559455
ipmi_sensor,server=10.20.2.203,unit=feet,name=altitude status=1i,value=80 1458488465012688613
ipmi_sensor,server=10.20.2.203,unit=watts,name=avg_power status=1i,value=220 1458488465012776511
ipmi_sensor,server=10.20.2.203,unit=volts,name=planar_3.3v status=1i,value=3.28 1458488465012861875
ipmi_sensor,server=10.20.2.203,unit=volts,name=planar_vbat status=1i,value=3.04 1458488465013072508
ipmi_sensor,server=10.20.2.203,unit=rpm,name=fan_1a_tach status=1i,value=2610 1458488465013137932
ipmi_sensor,server=10.20.2.203,unit=rpm,name=fan_1b_tach status=1i,value=1775 1458488465013279896
ipmi_sensor,server=10.20.2.203,name=uid_light,host=foo.bar.com value=0,status=1i 1517125513000000000
ipmi_sensor,server=10.20.2.203,name=sys._health_led,host=foo.bar.com status=1i,value=0 1517125513000000000
ipmi_sensor,server=10.20.2.203,unit=watts,host=foo.bar.com,name=power_supply_1 status=1i,value=110 1517125513000000000
ipmi_sensor,server=10.20.2.203,host=foo.bar.com,name=power_supply_2,unit=watts status=1i,value=120 1517125513000000000
ipmi_sensor,server=10.20.2.203,name=power_supplies,host=foo.bar.com value=0,status=1i 1517125513000000000
ipmi_sensor,server=10.20.2.203,name=fan_1,unit=percent,host=foo.bar.com status=1i,value=43.12 1517125513000000000
```


When retrieving stats from the local machine (no server specified):
```
ipmi_sensor,name=uid_light,host=foo.bar.com value=0,status=1i 1517125513000000000
ipmi_sensor,name=sys._health_led,host=foo.bar.com status=1i,value=0 1517125513000000000
ipmi_sensor,unit=watts,host=foo.bar.com,name=power_supply_1 status=1i,value=110 1517125513000000000
ipmi_sensor,host=foo.bar.com,name=power_supply_2,unit=watts status=1i,value=120 1517125513000000000
ipmi_sensor,name=power_supplies,host=foo.bar.com value=0,status=1i 1517125513000000000
ipmi_sensor,name=fan_1,unit=percent,host=foo.bar.com status=1i,value=43.12 1517125513000000000
```

#### Version 2 Schema

When retrieving stats from the local machine (no server specified):
```
ipmi_sensor,unit=degrees_c,name=ambient_temp status=1i,value=20 1458488465012559455
ipmi_sensor,unit=feet,name=altitude status=1i,value=80 1458488465012688613
ipmi_sensor,unit=watts,name=avg_power status=1i,value=220 1458488465012776511
ipmi_sensor,unit=volts,name=planar_3.3v status=1i,value=3.28 1458488465012861875
ipmi_sensor,unit=volts,name=planar_vbat status=1i,value=3.04 1458488465013072508
ipmi_sensor,unit=rpm,name=fan_1a_tach status=1i,value=2610 1458488465013137932
ipmi_sensor,unit=rpm,name=fan_1b_tach status=1i,value=1775 1458488465013279896
ipmi_sensor,name=uid_light,entity_id=23.1,status_code=ok,status_desc=ok,host=foo.bar.com value=0 1517125474000000000
ipmi_sensor,host=foo.bar.com,name=sys._health_led,entity_id=23.2,status_code=ok,status_desc=ok value=0 1517125474000000000
ipmi_sensor,status_code=ok,unit=watts,status_desc=presence_detected,host=foo.bar.com,name=power_supply_1,entity_id=10.1 value=110 1517125474000000000
ipmi_sensor,host=foo.bar.com,name=power_supply_2,entity_id=10.2,status_code=ok,unit=watts,status_desc=presence_detected value=125 1517125474000000000
ipmi_sensor,name=power_supplies,entity_id=10.3,status_code=ok,status_desc=fully_redundant,host=foo.bar.com value=0 1517125474000000000
ipmi_sensor,unit=percent,status_desc=transition_to_running,host=foo.bar.com,name=fan_1,entity_id=7.1,status_code=ok value=43.12 1517125474000000000
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If you regen this output with Telegraf 1.7, it should sort the tags. You can leave out the host tag since it is added everywhere.

```
113 changes: 96 additions & 17 deletions plugins/inputs/ipmi_sensor/ipmi.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@ package ipmi_sensor
import (
"fmt"
"os/exec"
"regexp"
"strconv"
"strings"
"sync"
Expand All @@ -17,11 +18,13 @@ var (
execCommand = exec.Command // execCommand is used to mock commands in tests.
)

// Ipmi stores the configuration values for the ipmi_sensor input plugin
type Ipmi struct {
Path string
Privilege string
Servers []string
Timeout internal.Duration
Path string
Privilege string
Servers []string
Timeout internal.Duration
SchemaVersion int
}

var sampleConfig = `
Expand All @@ -46,16 +49,22 @@ var sampleConfig = `

## Timeout for the ipmitool command to complete
timeout = "20s"

## Schema Version: (Optional, defaults to version 1)
schemaVersion = 2
`

// SampleConfig returns the documentation about the sample configuration
func (m *Ipmi) SampleConfig() string {
return sampleConfig
}

// Description returns a basic description for the plugin functions
func (m *Ipmi) Description() string {
return "Read metrics from the bare metal servers via IPMI"
}

// Gather is the main execution function for the plugin
func (m *Ipmi) Gather(acc telegraf.Accumulator) error {
if len(m.Path) == 0 {
return fmt.Errorf("ipmitool not found: verify that ipmitool is installed and that ipmitool is in your PATH")
Expand Down Expand Up @@ -93,23 +102,32 @@ func (m *Ipmi) parse(acc telegraf.Accumulator, server string) error {
opts = conn.options()
}
opts = append(opts, "sdr")
if m.SchemaVersion == 2 {
opts = append(opts, "elist")
}
cmd := execCommand(m.Path, opts...)
out, err := internal.CombinedOutputTimeout(cmd, m.Timeout.Duration)
if err != nil {
return fmt.Errorf("failed to run command %s: %s - %s", strings.Join(cmd.Args, " "), err, string(out))
}
if m.SchemaVersion == 2 {
return parseV2(acc, hostname, string(out))
}
return parseV1(acc, hostname, string(out))
}

func parseV1(acc telegraf.Accumulator, hostname string, cmdOut string) error {
// each line will look something like
// Planar VBAT | 3.05 Volts | ok
lines := strings.Split(string(out), "\n")
lines := strings.Split(cmdOut, "\n")
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'll bet this is was an existing bug in Windows due to only allowing CR line endings. What would be ideal is if we switched to bufio.Scanner here.

for i := 0; i < len(lines); i++ {
vals := strings.Split(lines[i], "|")
if len(vals) != 3 {
ipmiFields := extractFieldsFromRegex(`^(?P<name>[^|]*)\|(?P<description>[^|]*)\|(?P<status_code>.*)`, lines[i])
if len(ipmiFields) != 3 {
continue
}

tags := map[string]string{
"name": transform(vals[0]),
"name": transform(ipmiFields["name"]),
}

// tag the server is we have one
Expand All @@ -118,18 +136,16 @@ func (m *Ipmi) parse(acc telegraf.Accumulator, server string) error {
}

fields := make(map[string]interface{})
if strings.EqualFold("ok", trim(vals[2])) {
if strings.EqualFold("ok", trim(ipmiFields["status_code"])) {
fields["status"] = 1
} else {
fields["status"] = 0
}

val1 := trim(vals[1])

if strings.Index(val1, " ") > 0 {
if strings.Index(ipmiFields["description"], " ") > 0 {
// split middle column into value and unit
valunit := strings.SplitN(val1, " ", 2)
fields["value"] = Atofloat(valunit[0])
valunit := strings.SplitN(ipmiFields["description"], " ", 2)
fields["value"] = aToFloat(valunit[0])
if len(valunit) > 1 {
tags["unit"] = transform(valunit[1])
}
Expand All @@ -143,13 +159,76 @@ func (m *Ipmi) parse(acc telegraf.Accumulator, server string) error {
return nil
}

func Atofloat(val string) float64 {
func parseV2(acc telegraf.Accumulator, hostname string, cmdOut string) error {
// each line will look something like
// CMOS Battery | 65h | ok | 7.1 |
// Temp | 0Eh | ok | 3.1 | 55 degrees C
// Drive 0 | A0h | ok | 7.1 | Drive Present
lines := strings.Split(cmdOut, "\n")
for i := 0; i < len(lines); i++ {
ipmiFields := extractFieldsFromRegex(`^(?P<name>[^|]*)\|[^|]+\|(?P<status_code>[^|]*)\|(?P<entity_id>[^|]*)\|(?:(?P<description>[^|]+))?`, lines[i])
if len(ipmiFields) < 3 || len(ipmiFields) > 4 {
continue
}

tags := map[string]string{
"name": transform(ipmiFields["name"]),
}

// tag the server is we have one
if hostname != "" {
tags["server"] = hostname
}
tags["entity_id"] = transform(ipmiFields["entity_id"])
tags["status_code"] = trim(ipmiFields["status_code"])
fields := make(map[string]interface{})
descriptionResults := extractFieldsFromRegex(`^(?P<analogValue>[0-9.]+)\s(?P<analogUnit>.*)|(?P<status>.+)|^$`, trim(ipmiFields["description"]))
// This is an analog value with a unit
if descriptionResults["analogValue"] != "" && len(descriptionResults["analogUnit"]) >= 1 {
fields["value"] = aToFloat(descriptionResults["analogValue"])
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think if the value cannot be converted to a float then we should skip the line.

// Some implementations add an extra status to their analog units
unitResults := extractFieldsFromRegex(`^(?P<realAnalogUnit>[^,]+)(?:,\s*(?P<statusDesc>.*))?`, descriptionResults["analogUnit"])
tags["unit"] = transform(unitResults["realAnalogUnit"])
if unitResults["statusDesc"] != "" {
tags["status_desc"] = transform(unitResults["statusDesc"])
}
} else {
// This is a status value
fields["value"] = 0.0
// Extended status descriptions aren't required, in which case for consistency re-use the status code
if descriptionResults["status"] != "" {
tags["status_desc"] = transform(descriptionResults["status"])
} else {
tags["status_desc"] = transform(ipmiFields["status_code"])
}
}

acc.AddFields("ipmi_sensor", fields, tags, time.Now())
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Call time.Now() once before the loop, then pass the same time in for all metrics. Usually doesn't matter due to timestamp rounding but it is still a good idea.

}

return nil
}

// extractFieldsFromRegex consumes a regex with named capture groups and returns a kvp map of strings with the results
func extractFieldsFromRegex(regex string, input string) map[string]string {
re := regexp.MustCompile(regex)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We should compile all the regex once at the package level, then reuse the compiled regex here.

submatches := re.FindStringSubmatch(input)
results := make(map[string]string)
for i, name := range re.SubexpNames() {
if name != input && name != "" && input != "" {
results[name] = trim(submatches[i])
}
}
return results
}

// aToFloat converts string representations of numbers to float64 values
func aToFloat(val string) float64 {
f, err := strconv.ParseFloat(val, 64)
if err != nil {
return 0.0
} else {
return f
}
return f
}

func trim(s string) string {
Expand Down
Loading