Debug Information Files
Debug information files allow Sentry to extract stack traces and provide more information about crash reports for most compiled platforms. Information stored in debug files includes original function names, paths to source files and line numbers, source code context, or the placement of variables in memory. Sentry can use some of this information and display it on the issue details page.
Each major platform uses different debug information files. We currently support the following formats:
- dSYM files for iOS, iPadOS, tvOS, watchOS, and macOS
- ELF symbols for Linux and Android (NDK)
- PDB files for Windows
- Breakpad symbols for all platforms
- WASM files for WebAssembly
- ProGuard mappings for Java and Android
Source maps, while also being debug information files, are handled differently in Sentry. For more information see Source Maps in sentry-cli.
Sentry requires access to debug information files of your application as well as system libraries to provide fully symbolicated crash reports. You can either upload your files to Sentry or put them on a compatible Symbol Server to be downloaded by Sentry when needed.
Debug information files can be managed on the Debug Files section in Project Settings. This page lists all uploaded files and allows to configure symbol servers for automatic downloads.
Sentry differentiates in four kinds of debug information:
Debug Information: Provides function names, paths to source files, line numbers and inline frames. The process of resolving this information from instruction addresses is called "symbolication". This information is relatively large compared to the executable and usually put into a separate file. In Sentry, these files are designated as debug companions and show the
Symbol Tables: If debug information is not available for a certain library, Sentry can use symbol tables as a fallback to retrieve function names. Symbol tables are usually included in both the executable and debug companion files. However, they do not contain sufficient information to resolve inline functions or file names and line numbers. The
symtabtag indicates symbol tables.
Source Code: Conventionally, source code is not part of regular debug information files. Sentry CLI can bundle source code of your application and upload it to display source context in stack traces in Sentry. These bundles show up with the
Unwind Information: Enables Sentry to extract stack traces from Minidumps and other binary crash formats of optimized builds. This process is referred to as "stack unwinding" or "stack walking". Since this is also required when throwing exceptions in C++, this information is often included in the executable or library. If an uploaded file contains this information, it shows the
unwindtag. Note that on some platforms no actual unwinding takes place. For instance, WebAssembly currently does not have the equivalent of minidumps which means we do not require that type of information in such cases.
Compilers place the above debug information in different files passed on the target platform, architecture, build flags or optimization level. Consequently, Sentry might not need all of the above information to process crash reports. Still, it is always a good idea to provide all available debug information.
sentry-cli can be used to list properties of supported debug files and
validate their contents. See Debug Information Files in sentry-cli for more information.
The remainder of this section describes the file formats in detail.
MachO and dSYM
Executables, dynamic libraries and debug companions on all Apple platforms use the Mach Object, or short MachO, container format. This applies to iOS, iPadOS, tvOS, watchOS, and macOS.
Executables do not carry a file extension. For desktop applications, they are often placed in app bundle structures with the
.appsuffix. Unless stripped manually, executables contain unwind information and a symbol table. Debug information is never stored in executables.
Dynamic Libraries use the
.dylibextension, and otherwise behave exactly the same as executables.
Debug Companions are placed in a folder structure with the
.dSYMextension, and are located at
<name>.dSYM/Contents/Resources/DWARF/<name>. They usually contain a symbol table and debug information, but rarely unwind information.
When building an application with Xcode or with the
clang compiler, debug
information is automatically placed in a dSYM file. When linking manually,
however, the dSYM file must be created using the following command:
Executable and Linkable Format (ELF)
On Linux distributions, executables and debug information are stored in ELF containers. Unlike other platforms, there is no dedicated container or specifier for debug companion files.
Debug information is part of the binary (executable or library) and stripped
when generating release builds due to their size. However, there is a way to
retain them in a separate file (either in a different location or with
# There is an executable called "binary" in the CWD objcopy --only-keep-debug binary binary.debug objcopy --strip-debug --strip-unneeded binary objcopy --add-gnu-debuglink=binary.debug binary
This results in the following structure:
Executables do not carry a file extension. If stripped like above, executables contain a symbol table, but no debug information. If the build run omits frame pointers, unwind information will also be retained. Both can be further stripped using flags like
Shared Libraries use the
.soextension, and otherwise behave exactly the same as executables.
Debug Companions do not carry a standard file extension, but are often named
.debug. If stripped like above, these files contain unwind information, debug information and a symbol table.
Shared libraries installed via package managers usually provide their debugging
information in separate
*-dev packages and put it in locations like
/usr/local/debug/.... To receive symbolicated stack traces from those
libraries, make sure to also upload their symbols in addition to your app’s
ELF supports the compression of debug information which can significantly reduce
the time required to upload debug information files to Sentry and thus improve
gcc (version 5 or newer) and
clang (version 5 or newer) support
this by passing the
-gz flag to both the compiler and linker. The common way
to compress debug information, however, is when stripping the executable:
# Note the --compress-debug-sections option objcopy --only-keep-debug --compress-debug-sections=zlib binary binary.debug
This can be verified by checking for the
C flag in
readelf, corresponding to
readelf -S path/to/file ...  .debug_info PROGBITS 0000000000000000 00000370 000000000000e133 0000000000000000 C 0 0 1
PE and PDB
On Microsoft Windows, executables and dynamic libraries use the Portable Executable container, in short PE. Debug information is stored in Program Database files, commonly referred to as PDB.
Executables use an
.exefile extension. Only when compiled for 64-bit architectures, they contain unwind information. Otherwise, they do not contain any usable information and will be omitted for the upload to Sentry.
Dynamic Libraries use the
.dllfile extension, and otherwise behave exactly the same as executables.
Debug Companions are stored in
.pdbfiles. They usually contain debug information and in most cases symbol tables. For 32-bit programs, they also contain unwind information. In rare cases, they might have different file names than their corresponding executable.
At the moment, Sentry supports Native PDBs only. PDBs for the .NET platform are not supported.
The Google Breakpad library has established a platform-independent ASCII format to store debug information. Such files are usually generated for applications using Breakpad, Crashpad or the Electron Framework.
The Breakpad repository includes
dump_syms tools for each platform that can
convert from native debug files to Breakpad symbols. These converters bundle all
available information into one file, such that only one file needs to be
In contrast to native debug files, Breakpad symbols discard a lot of information that is not required to process minidumps. Most notably, inline functions are not declared, such that Sentry is not able to display inline frames in stack traces.
For WebAssembly, we support DWARF in WASM containers. Note that we do not support source maps, which are also a format used for WASM debugging, but have shortcomings that make them impractical for a crash reporting tool like Sentry.
Since WASM does not specify debug/build IDs yet, we provide a separate tool to add build IDs and split files called wasm-split to help you create a debug companion file ready for uploading to Sentry while removing all debug information from the release binary.
ProGuard mapping files allow Sentry to resolve obfuscated Java classpaths and method names into their original form. In that sense, they act as debug information files for Java and Android applications.
Each debug information file specifies a unique identifier. Crash reports declare these identifiers to allow debuggers and crash reporting systems to resolve the correct files. Sentry distinguishes two kinds of identifiers:
Code Identifier: The unique identifier of the executable or dynamic library -- the code file. The contents of this identifier are platform-dependent: MachO files use a UUID, ELF files a SHA hash, PE files use a concatenation of certain header attributes. For WebAssembly we use an embedded UUID in the
build_idsection of the file.
Debug Identifier: The unique identifier of the debug companion file. In contrast to the code identifier, Sentry enforces the same structure on all platforms. On Windows, this is the actual unique id of the PDB file; on all other platforms this is a lossy transformation of the code identifier.
When uploading debug information files to Sentry, the CLI and server will always compute a Debug Identifier for each uploaded file. This identifier is associated with executables and libraries as well as debug companions to ensure that they can be uniquely located via one common mechanism.
Debug information does not have to be associated with releases. The unique debug identifier ensures that Sentry can choose the right files for every crash report. However, it is still recommended to configure releases in the client to benefit from other features.
For native events, the issue details page displays a list of Loaded Images. This list contains the executable and all loaded dynamic libraries including their debug identifiers. You can copy this identifier and search for the exact files that match it in the Debug Files settings screen.
sentry-cli can help to print properties of debug information files like their
debug identifier. See Checking Debug Information Files for more information.
GNU Build Identifiers
For ELF files on Linux, Sentry uses the GNU build identifier to compute the
debug identifier. All recent compilers and linkers support the emission of build
IDs, but sometimes they might require additional configuration.
gcc does this
by default, for
clang use one of the following flags:
--build-id=uuidfor a fast but non-reproducible random identifier.
--build-id=sha1for a slower but reproducible identifier generated by hashing the first page of the code section.
The identifier needs to be present and identical in the binary as well
as stripped debug information files. If the ID is missing for some reason,
upload the files before stripping so that
sentry-cli can compute a stable
identifier from the unstripped file.
PDB Age Mismatches
Microsoft PDBs compose their identifiers from two parts: A unique signature and an age field. The signature is generated when the PDB is written initially and usually changes with every build. The age is a counter that is incremented every time the PDB is modified.
PE files, such as executables and dynamic libraries, specify the full identifier of the corresponding PDB in their header. This includes the age. If the PDB is modified after the PE has been generated, however, its age might diverge. This can lead to different identifiers:
PE: 3003763b-afcb-4a97-aae3-28de8f188d7c-2 PDB: 3003763b-afcb-4a97-aae3-28de8f188d7c-4
sentry-cli can detect these differences during the upload process and
associates the same identifier to both files. However, this requires that both
files are uploaded in the same invocation of the upload command. Otherwise, the
identifiers diverge and Sentry might not be able to resolve the correct file
WASM Build IDs
WebAssembly does not yet support build IDs. The option proposed to implement build IDs for WebAssembly (Build ID Section for WASM) has not yet found widespread adoption. Instead, we use a custom extension to WebAssembly.
Our recommendation is to embed a UUID in the
build_id custom section as
raw binary. Our
tool can do this for you automatically.
Unlike other debug information files, ProGuard files do not have an intrinsic
unique identifier. Sentry CLI assigns them a SHA1 UUID based on the checksum of
the file. You can use
sentry-cli difutil check on a ProGuard file to see the
If you need to generate the UUID yourself, you can do so with the following algorithm (Python code for reference):
import uuid NAMESPACE = uuid.uuid5(uuid.NAMESPACE_DNS, "guardsquare.com") def get_proguard_uuid(filename): with open(filename, 'rb') as f: return uuid.uuid5(NAMESPACE, f.read())
The most straightforward way to provide Sentry with debug information file is
to upload them using
sentry-cli. Depending on your workflow, you may want to
upload as part of your build pipeline or when deploying and publishing your
Files can be uploaded using the
upload-dif command. This command will scan a
given folder recursively for files and upload them to Sentry:
sentry-cli upload-dif -o <org> -p <project> /path/to/myfile.debug.wasm > Found 2 debug information files > Prepared debug information files for upload > Uploaded 2 missing debug information files > File processing complete: PENDING 1ddb3423-950a-3646-b17b-d4360e6acfc9 (MyApp; x86_64 executable) PENDING 1ddb3423-950a-3646-b17b-d4360e6acfc9 (MyApp; x86_64 debug companion)
For all available options and more information refer to Uploading Debug Information.
Always ensure that debug files are uploaded before deploying or releasing your
application so that crash reports can be processed. For manual testing, use the
sentry-cli upload-dif --wait before sending the first native crash or error
If you upload a file that was previously reported as missing by Sentry in a crash report, it may take up to one hour until Sentry starts using this file for new crash reports. Existing events and issues are not processed again.
In Sentry, your uploaded files are associated with projects. You can view and manage uploads at Project Settings > Debug Files. This screen displays the most important properties of debug files:
- The debug identifier. There may be multiple entries sharing the same identifier, if debug information is split across multiple files.
- The name of the debug file. Sentry uses the name on the file system when
uploading the file using
- The architecture, and type of the file. This can be used to distinguish files into executables and debug companions if their debug identifiers and names match.
- The debug information available in these files. This can include
- Metadata like the size and time of upload.
If multiple projects require the same file, it needs to be re-uploaded. Sentry does not access debug files stored in other projects to symbolicate crash reports.
Sentry can suspend incoming crash reports until all required debug information files have been uploaded. This feature is called Reprocessing. It can be configured in Project Settings > Processing Issues. By default, this feature is disabled.
If enabled, crash reports with missing debug files will not be displayed in the issues stream. Instead, you will receive a warning that events cannot be processed until all debug files have been uploaded.
Once an issue is shown in the issues stream, it is no longer processed. Even with enabled reprocessing, new file uploads will not effect such events.
At the moment, this feature only applies to iOS crashes sent with the Cocoa SDK and is not compatible with Symbol Servers.
Sentry can download debug information files from external repositories. This allows you to stop uploading debug files and instead configure a public symbol server or run your own. It is also possible to configure external repositories and upload debug files at the same time.
To configure external repositories, go to Project Settings > Debug Files. Above the list of uploaded files, there are two settings to configure external repositories:
Custom Repositories: Configures custom repositories containing debug files. You can choose from configuring an HTTP symbol server, Amazon S3 bucket or Google Cloud Storage bucket. This requires a Business or Enterprise plan.
Built-In Repositories: Allows to select from a list of pre-configured symbol servers. By default, iOS and Microsoft are enabled.
Sentry queries external repositories for debug information files in the order they are configured. If custom repositories are configured, those are probed first. Only debug information files that are not found on one of the custom repositories are queried from the built-in ones.
To enable a built-in repository, select it from the dropdown list. This immediately adds the repository and uses its debug information files to symbolicate new crash reports. Likewise, any built-in repository can be disabled by clicking on the X next to the name.
Adding or removing external repositories applies immediately. As a result, events may group differently with the new information and create new issues. Beware that these cause notifications to your team members.
Custom repositories are available for organizations on the Business and Enterprise plans.
Independent of the internal format, Sentry supports three kinds of custom repositories:
HTTP Symbol Server: An HTTP server that serves debug files at a configurable path. Lookups in the server should generally be case-insensitive, although an explicit casing can be configured in the settings.
Amazon S3 Bucket: Either an entire S3 bucket or a subdirectory. This requires
s3:GetObject, and optionally
s3:ListBucketpermissions for the configured Access Key. Lookups in the bucket are case sensitive, which is why we recommend storing all files lower-cased.
Google Cloud Storage Bucket: Either an entire GCS bucket or a subdirectory. This requires
storage.objects.listpermissions for the configured service account. Lookups in the bucket are case sensitive, which is why we recommend storing all files lower-cased.
Apart from authentication configuration, all types have common config parameters:
Name: A name to identify the repository.
Path Casing: Overrides which casing Sentry uses to query for debug information files. The default is a mixed case, which will use the case described in the next section. When overridden, all access is either lowercased or uppercased. Defaults to "mixed case".
Directory Layout: The internal structure of the bucket, or the protocol of the symbol server. There are three layouts to choose from which are discussed in the next section. Defaults to "Platform Specific".
Sentry supports multiple layouts for external repositories. Based on the selected layout and the file type, we try to download files at specific paths.
The following table contains a mapping from the supported layouts to file path schemas applied for specific files:
|Microsoft SymStore (index2.txt)||-||-||Index2||Index2||-|
The path schemas in the table above are defined as follows:
Breakpad always uses a Breakpad ID to store symbols. These identifiers can be computed from Debug Identifiers by removing dashes and applying the following casing rules:
- The signature part of the id (first 32 characters) are uppercase.
- The age part of the id (remaining characters) are lowercase.
The name of the symbol file is platform dependent. On Windows, the file
.pdb) is replaced with
.sym. On all
other platforms, the
.sym extension is appended to the full file name
including potential extensions.
The LLDB debugger on macOS can read debug symbols from File Mapped UUID Directories. The UUID is broken up by splitting the first 20 hex digits into 4 character chunks, and a directory is created for each chunk. In the final directory, LLDB usually expects a symlink named by the last 12 hex digits, which it follows to the actual dSYM file.
This is not actually an LLVM feature. This is in fact a feature of
CoreFoundation and exclusively implemented on macOS on top of spotlight.
Spotlight indexes these paths and the private
is used by lldb to locate the symbols. macOS uses the symlinks of those
Since the executable or library shares the same UUID as the dSYM file, the
former are distinguished with a
The hex digits are uppercase, the app suffix is lowercase.
5E01/2A64/6CC5/36F1/9B4D/A0564049169B.app(executable or library)
GDB supports multiple lookup methods, depending on the way the debug info file is specified. Sentry uses the Build ID Method: Assuming that a GNU build ID note or section has been written to the ELF file, this specifies a unique identifier for the executable which is also retained in the debug file.
The GNU build ID is a variable-length binary string, usually consisting of a
20-byte SHA1 hash of the code section (
.text). The lookup path is
pp/nnnnnnnn.debug, where pp are the first 2 hex characters of the build ID
bit string, and nnnnnnnn are the rest of the hex string. To look up
.debug suffix is omitted.
b5/381a457906d279073822a5ceb24c4bfef94ddb(executable or library)
b5/381a457906d279073822a5ceb24c4bfef94ddb.debug(stripped debug file)
SSQP Key Conventions are an extension to the original Microsoft Symbol Server protocol for .NET. It specifies lookup paths for PE, PDB, MachO and ELF files. The case of all lookup paths is generally lowercase except for the age field of PDB identifiers which should be uppercase.
For MachO files and ELF files, SSQP specifies to use the same identifiers as used in the LLDB and GNU build id method, respectively. See the sections above for more information. This results in the following paths for all possible file types:
_.debug/elf-buildid-sym-<buildid>/_.debug(ELF debug file)
SSQP specifies an additional lookup method by SHA1 checksum over the file contents, commonly used for source file lookups. Sentry does not support this lookup method.
The public symbol server provided by Microsoft used to only host PDBs for the Windows platform. These use a signature-age debug identifier in addition to the file name to locate symbols. File paths are identical to SSQP, except for the default casing rules:
- Filenames are as given
- The signature and age of a PDB identifier are uppercase.
- The timestamp of a PE identifier is uppercase, but the size is lowercase.
Since the original Microsoft Symbol Server did not serve ELF or MachO files, we do not recommend using this convention for these types. However, Sentry will support the SSQP conventions with adapted casing rules when this layout is selected.
This layout is identical to SymStore, except that the first two characters of the file name are prepended to the path as an additional folder.
Compression of Debug Files
Sentry supports the following compression methods when downloading debug information files from external sources: Gzip, zlib (both with and without header), Zstandard, and Cabinet (CAB).
The convention on Microsoft's Symbol Server protocol is to store such files with
the last character of the file extension replaced with
_. A full example would
KERNEL32.dll/590285E9e0000/KERNEL32.dl_. This is not required on your own
repositories, as Sentry detects compression on all paths.
If Sentry has access to application source code, it can show snippets of code around the location of stack frames. Certain SDKs can resolve this source context automatically, such as the Python SDK, because they have access to unobfuscated source code at runtime.
To get source context for native applications, source code needs to be uploaded
alongside the debug information files. The recommended way to do this is by
sentry-cli. See Creating Source Bundles for more information.
Source bundles show up as regular debug files on the Debug Files settings
page. They are designated as "source bundle" and feature a
sources tag. To
match them with crash reports, they carry the same debug file as the respective
debug information file they have been created from.