segfault error on post request

[davidporfirio7]

I'm working on a project using Chatwoot, and I'm integrating it with Typebot. However, I'm encountering a segmentation fault (segfault) when making a POST request to Typebot’s API. This is crashing my workers intermittently.

The crash seems unrelated to payload size — responses are typically under 100KB, well below the threshold. Sometimes the full bot flow runs without issues; other times, it crashes on the first message.

Context

I'm using HTTParty to make the requests. Here’s the main file handling the API calls:

module Toolzzbot
  class ApiClient
    include HTTParty

    BASE_URL = "https://typebot.io/api/v1".freeze
    TIMEOUT = 10
    MAX_RESPONSE_SIZE = 500_000

    def initialize(toolzzbot_id)
      @toolzzbot_id = toolzzbot_id
    end

    def start_chat(custom_attributes)
      url = "#{BASE_URL}/typebots/#{@toolzzbot_id}/startChat"
      begin
        response = HTTParty.post(
          url,
          headers: { 'Content-Type' => 'application/json' },
          body: { prefilledVariables: custom_attributes }.to_json,
          timeout: TIMEOUT
        )
        return response.parsed_response if response.success? && response['sessionId'].present?
      rescue => e
        # Intentionally suppressed
      end
      nil
    end

    def continue_chat(session_id, message_content)
      url = "#{BASE_URL}/sessions/#{session_id}/continueChat"
      begin
        safe_message = message_content.to_s.encode('UTF-8', invalid: :replace, undef: :replace, replace: '')
        response = HTTParty.post(
          url,
          headers: { 'Content-Type' => 'application/json' },
          body: { message: safe_message }.to_json,
          timeout: TIMEOUT
        )
        return nil if response.body.bytesize > MAX_RESPONSE_SIZE

        if response.success?
          begin
            result = JSON.parse(response.body)
            return result if result.is_a?(Hash) && result['messages'].is_a?(Array)
          rescue JSON::ParserError
            return nil
          end
        elsif response.code == 404 && response.parsed_response['code'] == 'NOT_FOUND'
          return { session_ended: true }
        end
      rescue => e
        # Intentionally suppressed
      end
      nil
    end
  end
end

Example Payloads

Start Chat

{
  "sessionId": "cmay0d2sk000z92rbdmtu3gbd",
  "resultId": "y4gf6be2icb1dn7wamhbommr",
  "typebot": {
    "id": "cmawjeigv000d7ko5y5hp54gt",
    "theme": {},
    "settings": {}
  },
  "messages": [{
    "id": "hufnw5m93wgy7bu80l3txlld",
    "type": "text",
    "content": {
      "richText": [{
        "type": "p",
        "children": [{
          "text": "olá"
        }]
      }]
    }
  }],
  "input": {
    "id": "h4u6ev48zicwim2lpgo5pwu9",
    "outgoingEdgeId": "reskt5boa39wtytnmuw1g69n",
    "type": "text input"
  }
}

Continue Chat

{
  "messages": [{
    "id": "brai0fhcax04o942reqi6iv2",
    "type": "text",
    "content": {
      "richText": [{
        "type": "p",
        "children": [{
          "text": "tudo bem?"
        }]
      }]
    }
  }],
  "input": {
    "id": "vt63oex2xa5rh40j7j85ukgo",
    "outgoingEdgeId": "bqleirc2ha6t98n7cndh9zgs",
    "type": "text input"
  }
}

Issues Observed

• Workers crash with a segfault, but logs don’t show much detail.
• Crashes are inconsistent – sometimes the same payload works fine.
• Response size is small, and I've added guards for large payloads and encoding issues.

---

If anyone has insights on what could be causing a segfault in this context — possibly a memory issue with the JSON parsing or HTTParty internals — I’d appreciate the help!

[jnunemaker]

What version of Ruby and the gem are you using?

…

On Wed, May 21, 2025 at 5:54 PM David ***@***.***> wrote: *davidporfirio7* created an issue (jnunemaker/httparty#814) <#814> I'm working on a project (using chatwoot) and trying to integrate it with typebot and when trying to create a post request I'm getting a segfault error what is crashing my workers the response barely gets to 100kb and doesn't seems to have a reason for crashing, sometimes I can execute the whole bot flow and others it crashes in the first message here is the main file which handles the api calls: `module Toolzzbot class ApiClient include HTTParty BASE_URL = "https://typebot.io/api/v1".freeze TIMEOUT = 10 # seconds MAX_RESPONSE_SIZE = 500_000 # 500KB limit def initialize(toolzzbot_id) @toolzzbot_id = toolzzbot_id end def start_chat(custom_attributes) url = ***@***.***_id}/startChat" begin response = HTTParty.post( url, headers: { 'Content-Type' => 'application/json' }, body: { prefilledVariables: custom_attributes }.to_json, timeout: TIMEOUT ) if response.success? && response['sessionId'].present? return response.parsed_response else return nil end rescue => e return nil end end def continue_chat(session_id, message_content) url = "#{BASE_URL}/sessions/#{session_id}/continueChat" begin # Ensure message content is properly encoded safe_message = message_content.to_s.encode('UTF-8', invalid: :replace, undef: :replace, replace: '') response = HTTParty.post( url, headers: { 'Content-Type' => 'application/json' }, body: { message: safe_message }.to_json, timeout: TIMEOUT ) # Check response size to prevent memory issues if response.body.bytesize > MAX_RESPONSE_SIZE return nil end if response.success? # Parse response in a controlled way begin result = JSON.parse(response.body) # Validate essential structure unless result.is_a?(Hash) && result['messages'].is_a?(Array) return nil end return result rescue JSON::ParserError => e return nil end elsif response.code == 404 && response.parsed_response['code'] == 'NOT_FOUND' return { session_ended: true } else return nil end rescue => e return nil end end end end` here is an example of the api calls: startchat { "sessionId": "cmay0d2sk000z92rbdmtu3gbd", "resultId": "y4gf6be2icb1dn7wamhbommr", "typebot": { "id": "cmawjeigv000d7ko5y5hp54gt", "theme": {}, "settings": {} }, "messages": [{ "id": "hufnw5m93wgy7bu80l3txlld", "type": "text", "content": { "richText": [{ "type": "p", "children": [{ "text": "olá" }] }] } }], "input": { "id": "h4u6ev48zicwim2lpgo5pwu9", "outgoingEdgeId": "reskt5boa39wtytnmuw1g69n", "type": "text input" } } continuechat { "messages": [{ "id": "brai0fhcax04o942reqi6iv2", "type": "text", "content": { "richText": [{ "type": "p", "children": [{ "text": "tudo bem?" }] }] } }], "input": { "id": "vt63oex2xa5rh40j7j85ukgo", "outgoingEdgeId": "bqleirc2ha6t98n7cndh9zgs", "type": "text input" } } image.png (view on web) <https://github.com/user-attachments/assets/d82d132a-966e-4547-9573-d95341505263> — Reply to this email directly, view it on GitHub <#814>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAAB22YOZZCYGQF6CHEIT327TY2RAVCNFSM6AAAAAB5UNFOUCVHI2DSMVQWIX3LMV43ASLTON2WKOZTGA4DCNBZHA3DEMI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[davidporfirio7]

What version of Ruby and the gem are you using?
…

On Wed, May 21, 2025 at 5:54 PM David @.***> wrote:
davidporfirio7 created an issue (jnunemaker/httparty#814)
<#814>

I'm working on a project (using chatwoot) and trying to integrate it with
typebot and when trying to create a post request I'm getting a segfault
error what is crashing my workers the response barely gets to 100kb and
doesn't seems to have a reason for crashing, sometimes I can execute the
whole bot flow and others it crashes in the first message here is the main
file which handles the api calls:

`module Toolzzbot
class ApiClient
include HTTParty

BASE_URL = "https://typebot.io/api/v1".freeze
TIMEOUT = 10 # seconds
MAX_RESPONSE_SIZE = 500_000 # 500KB limit

def initialize(toolzzbot_id)
@toolzzbot_id = toolzzbot_id
end

def start_chat(custom_attributes)
url = @.***_id}/startChat"

begin
response = HTTParty.post(
url,
headers: { 'Content-Type' => 'application/json' },
body: { prefilledVariables: custom_attributes }.to_json,
timeout: TIMEOUT
)

 if response.success? && response['sessionId'].present?
   return response.parsed_response
 else
   return nil
 end

rescue => e
return nil
end
end

def continue_chat(session_id, message_content)
url = "#{BASE_URL}/sessions/#{session_id}/continueChat"

begin
# Ensure message content is properly encoded
safe_message = message_content.to_s.encode('UTF-8', invalid: :replace, undef: :replace, replace: '')

 response = HTTParty.post(
   url,
   headers: { 'Content-Type' => 'application/json' },
   body: { message: safe_message }.to_json,
   timeout: TIMEOUT
 )

 # Check response size to prevent memory issues
 if response.body.bytesize > MAX_RESPONSE_SIZE
   return nil
 end

 if response.success?
   # Parse response in a controlled way
   begin
     result = JSON.parse(response.body)

     # Validate essential structure
     unless result.is_a?(Hash) && result['messages'].is_a?(Array)
       return nil
     end

     return result
   rescue JSON::ParserError => e
     return nil
   end
 elsif response.code == 404 && response.parsed_response['code'] == 'NOT_FOUND'
   return { session_ended: true }
 else
   return nil
 end

rescue => e
return nil
end
end

end
end`

here is an example of the api calls:

startchat
{ "sessionId": "cmay0d2sk000z92rbdmtu3gbd", "resultId":
"y4gf6be2icb1dn7wamhbommr", "typebot": { "id": "cmawjeigv000d7ko5y5hp54gt",
"theme": {}, "settings": {} }, "messages": [{ "id":
"hufnw5m93wgy7bu80l3txlld", "type": "text", "content": { "richText": [{
"type": "p", "children": [{ "text": "olá" }] }] } }], "input": { "id":
"h4u6ev48zicwim2lpgo5pwu9", "outgoingEdgeId": "reskt5boa39wtytnmuw1g69n",
"type": "text input" } }

continuechat
{ "messages": [{ "id": "brai0fhcax04o942reqi6iv2", "type": "text",
"content": { "richText": [{ "type": "p", "children": [{ "text": "tudo bem?"
}] }] } }], "input": { "id": "vt63oex2xa5rh40j7j85ukgo", "outgoingEdgeId":
"bqleirc2ha6t98n7cndh9zgs", "type": "text input" } }

image.png (view on web)
https://github.com/user-attachments/assets/d82d132a-966e-4547-9573-d95341505263

—
Reply to this email directly, view it on GitHub
<#814>, or unsubscribe
https://github.com/notifications/unsubscribe-auth/AAAAB22YOZZCYGQF6CHEIT327TY2RAVCNFSM6AAAAAB5UNFOUCVHI2DSMVQWIX3LMV43ASLTON2WKOZTGA4DCNBZHA3DEMI
.
You are receiving this because you are subscribed to this thread.Message
ID: @.***>

I'm using Ruby 3.2.2 and httparty 0.21.0

[jnunemaker]

Do you have any stack trace you can share?

…

On May 21, 2025 at 5:54:48 PM, David ***@***.***> wrote: *davidporfirio7* created an issue (jnunemaker/httparty#814) <#814> I'm working on a project (using chatwoot) and trying to integrate it with typebot and when trying to create a post request I'm getting a segfault error what is crashing my workers the response barely gets to 100kb and doesn't seems to have a reason for crashing, sometimes I can execute the whole bot flow and others it crashes in the first message here is the main file which handles the api calls: `module Toolzzbot class ApiClient include HTTParty BASE_URL = "https://typebot.io/api/v1".freeze TIMEOUT = 10 # seconds MAX_RESPONSE_SIZE = 500_000 # 500KB limit def initialize(toolzzbot_id) @toolzzbot_id = toolzzbot_id end def start_chat(custom_attributes) url = ***@***.***_id}/startChat" begin response = HTTParty.post( url, headers: { 'Content-Type' => 'application/json' }, body: { prefilledVariables: custom_attributes }.to_json, timeout: TIMEOUT ) if response.success? && response['sessionId'].present? return response.parsed_response else return nil end rescue => e return nil end end def continue_chat(session_id, message_content) url = "#{BASE_URL}/sessions/#{session_id}/continueChat" begin # Ensure message content is properly encoded safe_message = message_content.to_s.encode('UTF-8', invalid: :replace, undef: :replace, replace: '') response = HTTParty.post( url, headers: { 'Content-Type' => 'application/json' }, body: { message: safe_message }.to_json, timeout: TIMEOUT ) # Check response size to prevent memory issues if response.body.bytesize > MAX_RESPONSE_SIZE return nil end if response.success? # Parse response in a controlled way begin result = JSON.parse(response.body) # Validate essential structure unless result.is_a?(Hash) && result['messages'].is_a?(Array) return nil end return result rescue JSON::ParserError => e return nil end elsif response.code == 404 && response.parsed_response['code'] == 'NOT_FOUND' return { session_ended: true } else return nil end rescue => e return nil end end end end` here is an example of the api calls: startchat { "sessionId": "cmay0d2sk000z92rbdmtu3gbd", "resultId": "y4gf6be2icb1dn7wamhbommr", "typebot": { "id": "cmawjeigv000d7ko5y5hp54gt", "theme": {}, "settings": {} }, "messages": [{ "id": "hufnw5m93wgy7bu80l3txlld", "type": "text", "content": { "richText": [{ "type": "p", "children": [{ "text": "olá" }] }] } }], "input": { "id": "h4u6ev48zicwim2lpgo5pwu9", "outgoingEdgeId": "reskt5boa39wtytnmuw1g69n", "type": "text input" } } continuechat { "messages": [{ "id": "brai0fhcax04o942reqi6iv2", "type": "text", "content": { "richText": [{ "type": "p", "children": [{ "text": "tudo bem?" }] }] } }], "input": { "id": "vt63oex2xa5rh40j7j85ukgo", "outgoingEdgeId": "bqleirc2ha6t98n7cndh9zgs", "type": "text input" } } image.png (view on web) <https://github.com/user-attachments/assets/d82d132a-966e-4547-9573-d95341505263> — Reply to this email directly, view it on GitHub <#814>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAAB22YOZZCYGQF6CHEIT327TY2RAVCNFSM6AAAAAB5UNFOUCVHI2DSMVQWIX3LMV43ASLTON2WKOZTGA4DCNBZHA3DEMI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[davidporfirio7]

Do you have any stack trace you can share?
…

On May 21, 2025 at 5:54:48 PM, David @.***> wrote:
davidporfirio7 created an issue (jnunemaker/httparty#814)
<#814>

I'm working on a project (using chatwoot) and trying to integrate it with
typebot and when trying to create a post request I'm getting a segfault
error what is crashing my workers the response barely gets to 100kb and
doesn't seems to have a reason for crashing, sometimes I can execute the
whole bot flow and others it crashes in the first message here is the main
file which handles the api calls:

`module Toolzzbot
class ApiClient
include HTTParty

BASE_URL = "https://typebot.io/api/v1".freeze
TIMEOUT = 10 # seconds
MAX_RESPONSE_SIZE = 500_000 # 500KB limit

def initialize(toolzzbot_id)
@toolzzbot_id = toolzzbot_id
end

def start_chat(custom_attributes)
url = @.***_id}/startChat"

begin
response = HTTParty.post(
url,
headers: { 'Content-Type' => 'application/json' },
body: { prefilledVariables: custom_attributes }.to_json,
timeout: TIMEOUT
)

 if response.success? && response['sessionId'].present?
   return response.parsed_response
 else
   return nil
 end

rescue => e
return nil
end
end

def continue_chat(session_id, message_content)
url = "#{BASE_URL}/sessions/#{session_id}/continueChat"

begin
# Ensure message content is properly encoded
safe_message = message_content.to_s.encode('UTF-8', invalid: :replace, undef: :replace, replace: '')

 response = HTTParty.post(
   url,
   headers: { 'Content-Type' => 'application/json' },
   body: { message: safe_message }.to_json,
   timeout: TIMEOUT
 )

 # Check response size to prevent memory issues
 if response.body.bytesize > MAX_RESPONSE_SIZE
   return nil
 end

 if response.success?
   # Parse response in a controlled way
   begin
     result = JSON.parse(response.body)

     # Validate essential structure
     unless result.is_a?(Hash) && result['messages'].is_a?(Array)
       return nil
     end

     return result
   rescue JSON::ParserError => e
     return nil
   end
 elsif response.code == 404 && response.parsed_response['code'] == 'NOT_FOUND'
   return { session_ended: true }
 else
   return nil
 end

rescue => e
return nil
end
end

end
end`

here is an example of the api calls:

startchat
{ "sessionId": "cmay0d2sk000z92rbdmtu3gbd", "resultId":
"y4gf6be2icb1dn7wamhbommr", "typebot": { "id": "cmawjeigv000d7ko5y5hp54gt",
"theme": {}, "settings": {} }, "messages": [{ "id":
"hufnw5m93wgy7bu80l3txlld", "type": "text", "content": { "richText": [{
"type": "p", "children": [{ "text": "olá" }] }] } }], "input": { "id":
"h4u6ev48zicwim2lpgo5pwu9", "outgoingEdgeId": "reskt5boa39wtytnmuw1g69n",
"type": "text input" } }

continuechat
{ "messages": [{ "id": "brai0fhcax04o942reqi6iv2", "type": "text",
"content": { "richText": [{ "type": "p", "children": [{ "text": "tudo bem?"
}] }] } }], "input": { "id": "vt63oex2xa5rh40j7j85ukgo", "outgoingEdgeId":
"bqleirc2ha6t98n7cndh9zgs", "type": "text input" } }

image.png (view on web)
https://github.com/user-attachments/assets/d82d132a-966e-4547-9573-d95341505263

—
Reply to this email directly, view it on GitHub
<#814>, or unsubscribe
https://github.com/notifications/unsubscribe-auth/AAAAB22YOZZCYGQF6CHEIT327TY2RAVCNFSM6AAAAAB5UNFOUCVHI2DSMVQWIX3LMV43ASLTON2WKOZTGA4DCNBZHA3DEMI
.
You are receiving this because you are subscribed to this thread.Message
ID: @.***>

[jnunemaker]

I think this is closed by #823. Hoping to ship new release soon with that included.