400 error when space is in value

[r0b0tb4c0n]

i get the following error when my window shade is partially open:
Status: 400, Error: Bad Request, Headers: [X-Influxdb-Error:metric parse error: expected field at 6:98: "windowShade,deviceName=Office\ Shade,deviceId=190,hubName=Home,unit=windowShade value="partially\ open",valueBinary=0i 1777933070012000000", Content-Length:444, Date:Mon, 04 May 2026 22:28:21 GMT, Content-Type:application/json], Data: null

i'm not using the advanced. Looking at the influx documentation:
https://docs.influxdata.com/influxdb/v1/write_protocols/line_protocol_reference/#special-characters
you only need to escape double quotes and backslash characters for string field values.

as a test i made this:

private String escapeFieldValueStringForInfluxDB(String str) {
    if (str == null) {
        return '"null"'
    }
    
    str = str.replaceAll("\\\\", "\\\\\\\\") // Escape backslashes
    str = str.replaceAll("\"", "\\\\\"")     // Escape double quotes
    
    return str
}

and then changed line 883 to this (showing line 882 and 883 below):

            // Everything else is a string.
            value = '"' + escapeFieldValueStringForInfluxDB(evt.value) + '"'

the errors stopped and data started flowing again. i was then able to check my db and the "partially open" value was there. I'd make a PR, but i'm not 100% sure that the change of line 883 is safe/won't cause unintended side-effects

[dennypage]

Interesting. I don't recall InfluxDB having an issue with escaping spaces in values previously. Also strange that the error is being marked on the first character following the space, as if the double quotes were not being processed correctly. Do you have debug log entries showing the exact text that was actually sent by InfluxDB-Logger?

Also, what version of the InfluxDB server are you using?

[r0b0tb4c0n]

for influxdb version, i have a weird setup. I am using telegraf as an proxy so that i can use tokens/ssl which i then use to relay that data into questdb. the error was on the telegraf ingest side before it went to relay to questdb. if you go straight into influxdb it seems to be more forgiving. historically i had been using influxdb2 and then upgraded to v3, but switched to questdb due to licensing limitations.

for a non-working example here is the hubitat debug log entry:

app:612026-05-05 09:08:40.637 AMwarnPost of 2 events failed - elapsed time 0.06 seconds - Status: 400, Error: Bad Request, Headers: [X-Influxdb-Error:metric parse error: expected field at 1:98: "windowShade,deviceName=Office\\ Shade,deviceId=190,hubName=Home,unit=windowShade value=\"partially\\ open\",valueBinary=0i 1777990105517000000", Content-Length:444, Date:Tue, 05 May 2026 14:08:40 GMT, Content-Type:application/json], Data: null
app:612026-05-05 09:08:40.575 AMdebugNumber of events queued for InfluxDB: 2
app:612026-05-05 09:08:25.636 AMinfoQueued event: level,deviceName=Office\ Shade,deviceId=190,hubName=Home,unit=% value=90 1777990105520000000
app:612026-05-05 09:08:25.633 AMdebugHandle Event: Office Shade, level, %, 90
app:612026-05-05 09:08:25.560 AMdebugScheduling batch
app:612026-05-05 09:08:25.559 AMinfoQueued event: windowShade,deviceName=Office\ Shade,deviceId=190,hubName=Home,unit=windowShade value="partially\ open",valueBinary=0i 1777990105517000000
app:612026-05-05 09:08:25.557 AMdebugHandle Event: Office Shade, windowShade, null, partially open

and here is the input log entry from the telegraf side:

[inputs.influxdb_v2_listener] Error parsing the request body: metric parse error: expected field at 47:98: "windowShade,deviceName=Office\\ Shade,deviceId=190,hubName=Home,unit=windowShade value=\"partially\\ open\",valueBinary=0i 1777934270011000000"

after the code change, there is no error in the input log. here is the hubitat debug log entry:

[app:61](https://10.10.10.130/logs?tab=past&appId=61#)2026-05-05 09:04:16.607 AMinfoQueued event: level,deviceName=Office\ Shade,deviceId=190,hubName=Home,unit=% value=91 1777989856484000000
[app:61](https://10.10.10.130/logs?tab=past&appId=61#)2026-05-05 09:04:16.604 AMdebugHandle Event: Office Shade, level, %, 91
[app:61](https://10.10.10.130/logs?tab=past&appId=61#)2026-05-05 09:04:16.510 AMinfoQueued event: windowShade,deviceName=Office\ Shade,deviceId=190,hubName=Home,unit=windowShade value="partially open",valueBinary=0i 1777989856479000000
[app:61](https://10.10.10.130/logs?tab=past&appId=61#)2026-05-05 09:04:16.507 AMdebugHandle Event: Office Shade, windowShade, null, partially open

and i see this in the output log:

windowShade,deviceId=190,deviceName=Office\ Shade,host=b353de71f41a,hubName=Home,unit=windowShade value="partially open",valueBinary=0i 1777989192220000000

as you can see that the value field is not escaped in the working version

as a side note, i also hit this error when i tried to send the formatted date from the hub information driver.

if you want to replicate on your side, here's my telegraf config:

[agent]
  debug = true
  logfile = "/logs/telegraf.log"
  logfile_rotation_max_size = "50MB"
  logfile_rotation_max_archives = 3
  interval = "10s"
  round_interval = true
  metric_batch_size = 1000
  metric_buffer_limit = 10000
  collection_jitter = "0s"
  flush_interval = "10s"
  flush_jitter = "0s"

[[inputs.influxdb_v2_listener]]
  service_address = ":8286"
  token = "redacted"
  tls_cert = "/etc/telegraf/cert.pem"
  tls_key = "/etc/telegraf/key.pem"
  tls_min_version = "TLS12"

[[processors.regex]]
  [[processors.regex.field_rename]]
    pattern = "[/-]"
    replacement = "_"
  [[processors.regex.tag_rename]]
    pattern = "[/-]"
    replacement = "_"
  [[processors.regex.metric_rename]]
    pattern = "[/-]"
    replacement = "_"

[[outputs.socket_writer]]
  address = "tcp://questdb:9009"
  data_format = "influx"
  keep_alive_period = "30s"

[[outputs.file]]
  files = ["/logs/telegraf_output.log"]
  data_format = "influx"
  rotation_max_size = "50MB"
  rotation_max_archives = 3

the regex processors are in there because i run another app that puts those symbols in the measurement which quest rejects for table names

[dennypage]

I've re-confirmed that the syntax is valid for InfluxDB. It looks like Telegraph's InfluxDB protocol parser is broken.

While I agree the v1.8 documentation indicates that escaping spaces in data values is not required, the parser does work as expected if you do. And not knowing how prior InfluxDB protocols behaved, I am rather reluctant to change InfluxDB-Logger to remove the escaping without some extensive testing.

[r0b0tb4c0n]

i'd be more than happy to do the testing, i can spin up instances of influx 1.x 2.x and 3.x pretty quick, i'd just want to know what you would want me to test. Would taking everything in hub information driver + all the attributes available for all devices in the advanced options page be enough or would there be more that should be done?